{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-harness-pick-2026-09-29-public-capture-identity", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 32, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 25, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 7, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 22, "population": "source-reported planned cells including comparison setup", "qualifier": "about"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 35, "population": "source-reported attempts including aborts", "qualifier": "about"}], "evidence": [{"path": "rounds/harness-pick-2026-09-29/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": ["codex__gpt-6-luna__high__default__hp-syn-13__r1-delivery-ef191f17d6dc", "codex__gpt-6-luna__high__default__hp-syn-23__r1-delivery-abf03c15a8a8", "codex__gpt-6-luna__high__default__hp-syn-40__r1-delivery-c61b8671583b", "codex__gpt-6-sol__high__default__hp-syn-40__r1-delivery-4785810cecdc", "grok__grok-4.7__high__default__hp-syn-13__r1-delivery-11554abc4a9c", "grok__grok-4.7__high__default__hp-syn-13__r2-delivery-bd9fcc306e22", "grok__grok-4.7__high__default__hp-syn-23__r1-delivery-b7daed315a8e", "grok__grok-4.7__high__default__hp-syn-23__r2-delivery-5ec9870f1686", "grok__grok-4.7__high__default__hp-syn-40__r1-delivery-1739bf9ee368", "grok__grok-4.7__high__default__hp-syn-40__r1-delivery-5346f1200d62", "grok__grok-4.7__high__default__hp-syn-40__r2-delivery-36242d5460c7", "kh__opencode-go-deepseek-v4-pro__max__default__hp-syn-23__r1-delivery-32e99f57a10d", "opencode__openai-gpt-6-luna__high__default__hp-syn-23__r1-delivery-6c9841349db3", "opencode__xai-grok-4.7__high__default__hp-syn-23__r1-delivery-53cc48bd31e3", "pi__openai-codex-gpt-6-luna__high__default__hp-syn-13__r1-delivery-110ddff7a83d", "pi__openai-codex-gpt-6-luna__high__default__hp-syn-23__r1-delivery-c3ea88c5c8c1", "pi__openai-codex-gpt-6-luna__high__default__hp-syn-40__r1-delivery-6c04ed6281df", "pi__openai-codex-gpt-6-sol__high__default__hp-syn-40__r1-delivery-febffe2e59cd", "pi__xai-grok-4.7__high__default__hp-syn-13__r1-delivery-279c8ab01ccd", "pi__xai-grok-4.7__high__default__hp-syn-13__r2-delivery-9f40acdbc112", "pi__xai-grok-4.7__high__default__hp-syn-23__r1-delivery-a4b34c49a11b", "pi__xai-grok-4.7__high__default__hp-syn-23__r2-delivery-ea1ed08f3879", "pi__xai-grok-4.7__high__default__hp-syn-40__r1-delivery-25e5d8412c05", "pi__xai-grok-4.7__high__default__hp-syn-40__r1-delivery-38975fc71bf4", "pi__xai-grok-4.7__high__default__hp-syn-40__r2-delivery-b1976379a3ff"], "hypothesis_ids": ["H19", "H20"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 25, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 32, "non_unique_handles": 7}, "round_id": "harness-pick-2026-09-29", "round_page": "rounds/harness-pick-2026-09-29/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-kh-compare-2026-09-29-public-capture-identity", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 33, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 32, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 1, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 3, "population": "source-reported scored design", "unit": "tasks at one repetition each"}], "evidence": [{"path": "rounds/kh-compare-2026-09-29/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": ["kh-compare__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r1-delivery-2189eddabedd", "kh-compare__gpt-6.1-sol__high__default__elx-02-ingest-supervision__r2-delivery-03ef516419a6", "kh-compare__gpt-6.1-sol__high__default__elx-06-command-runner__r1-delivery-999fba145738", "kh-compare__gpt-6.1-sol__high__default__elx-06-command-runner__r2-delivery-885b25ecd30c", "kh-compare__gpt-6.1-sol__high__default__new-beltway-durable-leases__r1-delivery-ac23d92f27bb", "kh-compare__gpt-6.1-sol__high__default__new-beltway-durable-leases__r2-delivery-02a08f7c1998", "kh-compare__gpt-6.1-sol__high__default__new-beltway-generation-cache__r1-delivery-5e83c5e8957b", "kh-compare__gpt-6.1-sol__high__default__new-beltway-generation-cache__r2-delivery-4f409ba5a1e4", "kh-compare__gpt-6.1-sol__high__default__new-beltway-pipeline-deadline__r1-delivery-e958b526e558", "kh-compare__gpt-6.1-sol__high__default__new-beltway-pipeline-deadline__r2-delivery-0d76730dac98", "kh-compare__gpt-6.1-sol__high__default__new-tally-external-merge__r1-delivery-dc682facb37f", "kh-compare__gpt-6.1-sol__high__default__new-tally-external-merge__r2-delivery-e5150a0fc40c", "kh-compare__gpt-6.1-sol__high__default__new-trackline-bulk-transfer-outbox__r1-delivery-e2097c0486e0", "kh-compare__gpt-6.1-sol__high__default__new-trackline-bulk-transfer-outbox__r2-delivery-5453d1f857c9", "kh-compare__gpt-6.1-sol__high__default__new-trackline-resumable-import__r1-delivery-5163bf333ef0", "kh-compare__gpt-6.1-sol__high__default__new-trackline-resumable-import__r2-delivery-2b7baa822eaf", "kh-compare__gpt-6.1-sol__high__default__rails-ft-activity-feed-api__r1-delivery-eba7f481b808", "kh-compare__gpt-6.1-sol__high__default__rails-ft-activity-feed-api__r2-delivery-fbae885b376e", "kh-compare__gpt-6.1-sol__high__default__rails-ft-board-publish-unpublish-public-boundary__r1-delivery-dbc86e1351fc", "kh-compare__gpt-6.1-sol__high__default__rails-ft-board-publish-unpublish-public-boundary__r2-delivery-8b54e8ad7e63", "kh-compare__gpt-6.1-sol__high__default__rails-ft-cancelled-account-cleanup__r1-delivery-f7a1969e58dc", "kh-compare__gpt-6.1-sol__high__default__rails-ft-cancelled-account-cleanup__r2-delivery-9cd1cb7292d6", "kh-compare__gpt-6.1-sol__high__default__rails-ft-comment-reaction-live-echo__r1-delivery-371c133a6833", "kh-compare__gpt-6.1-sol__high__default__rails-ft-comment-reaction-live-echo__r2-delivery-90f30d2af889", "kh-compare__gpt-6.1-sol__high__default__rt-08__r1-delivery-c6d3974611a9", "kh-compare__gpt-6.1-sol__high__default__rt-08__r2-delivery-85790acbe188", "kh-compare__gpt-6.1-sol__high__default__syn-20-email-invite-flow__r1-delivery-b1a569d5f305", "kh-compare__gpt-6.1-sol__high__default__syn-20-email-invite-flow__r2-delivery-3c94556ee0d9", "kh-compare__gpt-6.1-sol__high__default__syn-24-csv-import__r1-delivery-3dc2367028bb", "kh-compare__gpt-6.1-sol__high__default__syn-24-csv-import__r2-delivery-09f250420ba0", "kh-compare__gpt-6.1-sol__high__default__syn-26-archive-audit-log__r1-delivery-59ee29737394", "kh-compare__gpt-6.1-sol__high__default__syn-26-archive-audit-log__r2-delivery-82dada1ae3d9"], "hypothesis_ids": ["H21", "H23", "H26", "H28"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 32, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 33, "non_unique_handles": 1}, "round_id": "kh-compare-2026-09-29", "round_page": "rounds/kh-compare-2026-09-29/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-claim-b-screen-1-public-capture-identity", "claim_status": "CONFOUNDED", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 0, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 12, "population": "source-reported model cells", "unit": "cells"}], "evidence": [{"path": "rounds/claim-b-screen-1/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H22"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 0, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 0, "non_unique_handles": 0}, "round_id": "claim-b-screen-1", "round_page": "rounds/claim-b-screen-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-claim-b-rerun-1-public-capture-identity", "claim_status": "CONFOUNDED", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 0, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 36, "population": "source-reported model cells after setup amendment", "unit": "model cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "pre-model setup failures", "unit": "setup failures"}], "evidence": [{"path": "rounds/claim-b-rerun-1/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H22"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 0, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 0, "non_unique_handles": 0}, "round_id": "claim-b-rerun-1", "round_page": "rounds/claim-b-rerun-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-confirm-combined-1-public-capture-identity", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 16, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 16, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 64, "population": "source-reported four-arm scored design", "unit": "cells"}], "evidence": [{"path": "rounds/confirm-combined-1/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": ["kh-confirm-combined__gpt-6.1-sol__high__default__elx-03-fanout-timeouts__r1-delivery-f1e34c717e80", "kh-confirm-combined__gpt-6.1-sol__high__default__elx-03-fanout-timeouts__r2-delivery-9d5e6537154b", "kh-confirm-combined__gpt-6.1-sol__high__default__elx-04-queue-backpressure__r1-delivery-e4089e8ac9f5", "kh-confirm-combined__gpt-6.1-sol__high__default__elx-04-queue-backpressure__r2-delivery-3a7b0a1cb8b4", "kh-confirm-combined__gpt-6.1-sol__high__default__rails-ar-tenant-isolation__r1-delivery-ea9732ebdeb5", "kh-confirm-combined__gpt-6.1-sol__high__default__rails-ar-tenant-isolation__r2-delivery-d4211ddd85ac", "kh-confirm-combined__gpt-6.1-sol__high__default__rails-ft-local-time-human-clock__r1-delivery-d3adee3e05b0", "kh-confirm-combined__gpt-6.1-sol__high__default__rails-ft-local-time-human-clock__r2-delivery-54bff5ee182e", "kh-confirm-combined__gpt-6.1-sol__high__default__rt-05__r1-delivery-73311e0773bd", "kh-confirm-combined__gpt-6.1-sol__high__default__rt-05__r2-delivery-fa7bff275587", "kh-confirm-combined__gpt-6.1-sol__high__default__rt-06__r1-delivery-8d85152ec3d8", "kh-confirm-combined__gpt-6.1-sol__high__default__rt-06__r2-delivery-540725d36725", "kh-confirm-combined__gpt-6.1-sol__high__default__syn-03-authz-cross-org-tickets__r1-delivery-f09f007993d4", "kh-confirm-combined__gpt-6.1-sol__high__default__syn-03-authz-cross-org-tickets__r2-delivery-1cec94d67a8e", "kh-confirm-combined__gpt-6.1-sol__high__default__syn-31-inbound-email-webhook__r1-delivery-0da4c0679505", "kh-confirm-combined__gpt-6.1-sol__high__default__syn-31-inbound-email-webhook__r2-delivery-26ce55716eab"], "hypothesis_ids": ["H21", "H22"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 16, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 16, "non_unique_handles": 0}, "round_id": "confirm-combined-1", "round_page": "rounds/confirm-combined-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-harness-compare-1-public-capture-identity", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 0, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 80, "population": "source-reported scored comparison", "unit": "graded cells"}], "evidence": [{"path": "rounds/harness-compare-1/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H22", "H28"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 0, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 0, "non_unique_handles": 0}, "round_id": "harness-compare-1", "round_page": "rounds/harness-compare-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact IDs, non-unique handles, and exact official-grade matches separately. Do not treat candidate deliveries as the historical scored cohort.", "arm": null, "claim_id": "CL-WA-hard1-calibrate-1-public-capture-identity", "claim_status": "CONFOUNDED", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 0, "population": "public candidate-manifest rows"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "unique public delivery IDs"}, {"evidence_status": "reproducible from public data", "n": 0, "population": "captured handles not accepted as unique cell IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 48, "population": "source-reported calibration", "unit": "cells"}], "evidence": [{"path": "rounds/hard1-calibrate-1/public-capture-manifest.csv", "selector": "all data rows; exactness from record_identity_status"}, {"path": "results/run-records/index.json", "selector": "manifest source_file and matching candidate IDs"}, {"path": "results/cells.jsonl", "selector": "official_grade_match field in candidate manifest"}], "evidence_status": "reproducible from public data for candidate capture counts; source-reported, not reproducible from public data for the historical design variants", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H22"], "interpretation": "The captured rows do not reconstruct the historical score cohort; no harness comparison is supported.", "metric": "candidate_delivery_capture", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"exact_delivery_ids": 0, "exact_official_grade_matches": 0, "manifest_has_arm_labels": false, "manifest_rows": 0, "non_unique_handles": 0}, "round_id": "hard1-calibrate-1", "round_page": "rounds/hard1-calibrate-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-baseline", "claim_status": "INTERIM", "comparison": null, "denominator": 21, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "reported_denominator"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 12, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "B0", "classification": "source-reported outcome; official public cell join unavailable", "experiment_id": "kh-climb-baseline", "planned_n": 21, "reported_denominator": 21, "reported_passes": 12}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-control", "claim_status": "INTERIM", "comparison": null, "denominator": 21, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "reported_denominator"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 14, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "unchanged H0", "classification": "source-reported outcome; two-pass difference described as noise floor", "experiment_id": "kh-climb-control", "planned_n": 21, "reported_denominator": 21, "reported_passes": 14}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-v1", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "V1 boundary prompt", "classification": "denominator conflict; do not calculate a contrast", "experiment_id": "kh-climb-v1", "planned_n": 21, "reported_passes": "conflicting: 10/18 usable; separately 10 pass + 11 fail + 0 infra / 21"}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-v2", "claim_status": "INTERIM", "comparison": null, "denominator": 21, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "reported_denominator"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 10, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "V2 mix-test", "classification": "source-reported outcome", "experiment_id": "kh-climb-v2", "planned_n": 21, "reported_denominator": 21, "reported_passes": 10}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-v3", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "started_n"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/4"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "V3", "classification": "built, not run", "experiment_id": "kh-climb-v3", "planned_n": null, "started_n": 0}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-v4", "claim_status": "INTERIM", "comparison": null, "denominator": 10, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 10, "population": "started_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 10, "population": "reported_denominator"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 11, "population": "unstarted_n"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/5"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 6, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "V4 verification gate", "classification": "partial; not final negative result", "experiment_id": "kh-climb-v4", "planned_n": 21, "reported_denominator": 10, "reported_passes": 6, "started_n": 10, "unstarted_n": 11}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-kh-climb-reported-kh-climb-v5", "claim_status": "INTERIM", "comparison": null, "denominator": 6, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 21, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "started_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "reported_denominator"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 15, "population": "unstarted_n"}], "evidence": [{"path": "rounds/kh-climb/reported-data.json", "pointer": "/experiment_records/6"}, {"path": "rounds/kh-climb/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H21", "H32", "H33"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 2, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm": "V5 requirement-to-diff", "classification": "partial; not final negative result", "experiment_id": "kh-climb-v5", "planned_n": 21, "reported_denominator": 6, "reported_passes": 2, "started_n": 6, "unstarted_n": 15}, "round_id": "kh-climb", "round_page": "rounds/kh-climb/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-ladder-reported-x-ladder-model-screen", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-ladder/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-ladder/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H41", "H42", "H47", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Luna max; Sol 6.1 high; Luna high", "classification": "source-reported small development screen", "experiment_id": "x-ladder-model-screen", "planned_n": "8 tasks × 2 reps for reported primary comparison", "reported_results": "15/16; 15/16; 12/16 respectively"}, "round_id": "x-ladder", "round_page": "rounds/x-ladder/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-ladder-reported-x-ladder-matrix", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-ladder/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-ladder/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H41", "H42", "H47", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "multiple ladder profiles", "classification": "separate source-reported matrix; arm crosswalk unavailable", "experiment_id": "x-ladder-matrix", "planned_n": null}, "round_id": "x-ladder", "round_page": "rounds/x-ladder/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-ladder-reported-x-ladder-time-first", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-ladder/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-ladder/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H41", "H42", "H47", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "time-first ladder vs control", "classification": "source-reported small screen", "experiment_id": "x-ladder-time-first", "reported_results": "6/6 vs 6/6; 2 rescues; about 141 additional seconds per attempt"}, "round_id": "x-ladder", "round_page": "rounds/x-ladder/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-ladder-reported-x-ladder-patient", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-ladder/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/x-ladder/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H41", "H42", "H47", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "patient profile vs control", "classification": "source-reported small screen", "experiment_id": "x-ladder-patient", "reported_results": "4/4 vs 4/4"}, "round_id": "x-ladder", "round_page": "rounds/x-ladder/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-ladder-reported-x-ladder-checkpoints", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-ladder/reported-data.json", "pointer": "/experiment_records/4"}, {"path": "rounds/x-ladder/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H41", "H42", "H47", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "reproducible checkpoint scan", "classification": "development inventory; not a task-success comparison", "experiment_id": "x-ladder-checkpoints", "reported_results": "0/145 runs"}, "round_id": "x-ladder", "round_page": "rounds/x-ladder/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-models-reported-x-models-overall", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 204, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 155, "population": "completed_graded_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 49, "population": "cancelled_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 122, "population": "correct_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 33, "population": "incorrect_n"}], "evidence": [{"path": "rounds/x-models/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-models/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H40", "H41", "H42", "H43", "H46"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "multiple models and efforts", "cancelled_n": 49, "classification": "source-reported inventory; cell-level rows unavailable", "completed_graded_n": 155, "correct_n": 122, "experiment_id": "x-models-overall", "incorrect_n": 33, "planned_n": 204}, "round_id": "x-models", "round_page": "rounds/x-models/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-models-reported-x-models-primary-screen", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 8, "population": "task_n"}], "evidence": [{"path": "rounds/x-models/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-models/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H40", "H41", "H42", "H43", "H46"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Luna max vs Sol 6.1 high", "classification": "source-reported inconclusive screen", "experiment_id": "x-models-primary-screen", "reported_results": "7/8 vs 6/8", "task_n": 8}, "round_id": "x-models", "round_page": "rounds/x-models/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-models-reported-x-models-adjacent-panel", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 8, "population": "task_n"}], "evidence": [{"path": "rounds/x-models/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-models/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H40", "H41", "H42", "H43", "H46"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Luna max; Sol 6.1 high; Luna high", "classification": "separate development panel", "experiment_id": "x-models-adjacent-panel", "reported_results": "15/16; 15/16; 12/16", "reps_per_arm": 2, "task_n": 8}, "round_id": "x-models", "round_page": "rounds/x-models/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-models-reported-x-models-astra", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-models/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/x-models/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H39", "H40", "H41", "H42", "H43", "H46"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Astra medium vs Sol high", "classification": "separate small screen; exact alignment unavailable", "experiment_id": "x-models-astra", "reported_results": "16/16"}, "round_id": "x-models", "round_page": "rounds/x-models/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-roles-reported-x-roles-worker-pilot", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 1, "population": "task_n"}], "evidence": [{"path": "rounds/x-roles/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-roles/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H39", "H43", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Sol high vs Luna max", "classification": "small pilot", "cost_wall": "about 7× lower API-equivalent cost and 1.7× wall for Luna, source-reported", "experiment_id": "x-roles-worker-pilot", "reported_results": "2/2 vs 2/2", "task_n": 1}, "round_id": "x-roles", "round_page": "rounds/x-roles/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-roles-reported-x-roles-implementation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 1, "population": "infrastructure_n"}], "evidence": [{"path": "rounds/x-roles/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-roles/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H39", "H43", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Sol high vs Luna high", "classification": "source-reported; cohort alignment not recovered", "experiment_id": "x-roles-implementation", "infrastructure_n": 1, "reported_results": "4/7 vs 2/7"}, "round_id": "x-roles", "round_page": "rounds/x-roles/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-roles-reported-x-roles-reviewer", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-roles/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-roles/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H39", "H43", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "Sol high vs Luna high", "classification": "small reviewer-envelope screen", "experiment_id": "x-roles-reviewer", "reported_results": "0/4 vs 2/2 complete"}, "round_id": "x-roles", "round_page": "rounds/x-roles/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-roles-reported-x-roles-decision", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-roles/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/x-roles/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H39", "H43", "H49"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "no promotion", "experiment_id": "x-roles-decision", "reported_results": "no role policy selected"}, "round_id": "x-roles", "round_page": "rounds/x-roles/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-review-reported-x-review-checklist", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 14, "population": "false_accept_denominator"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 10, "population": "false_block_denominator"}], "evidence": [{"path": "rounds/x-review/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-review/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H67", "H68", "H69"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "checklist and risk-directed review", "classification": "offline review-case diagnostic", "experiment_id": "x-review-checklist", "false_accept_denominator": 14, "false_accepts": 0, "false_block_denominator": 10, "false_blocks": 0, "supported_defect_recall": "14/14 each, source summary; case-level join unavailable"}, "round_id": "x-review", "round_page": "rounds/x-review/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-review-reported-x-review-repair", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 4, "population": "repair_sessions_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 1, "population": "required_final_emitted_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "fresh_confirmation_calls_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 8, "population": "fresh_confirmation_calls_expected_n"}], "evidence": [{"path": "rounds/x-review/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-review/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H67", "H68", "H69"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "partial repair workflow; not a task completion comparison", "experiment_id": "x-review-repair", "fresh_confirmation_calls_expected_n": 8, "fresh_confirmation_calls_n": 0, "repair_sessions_n": 4, "required_final_emitted_n": 1}, "round_id": "x-review", "round_page": "rounds/x-review/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-review-reported-x-review-semantic-repair", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-review/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-review/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H67", "H68", "H69"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate small source-reported screen; do not combine with review-case diagnostic", "experiment_id": "x-review-semantic-repair", "reported_result": "on complete delivery, 1/2 vs 0/2 failed the confirmation gate"}, "round_id": "x-review", "round_page": "rounds/x-review/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-race-reported-x-race", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "task_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 48, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 48, "population": "graded_n"}], "evidence": [{"path": "rounds/x-race/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-race/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": 48, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "four race/selection arms", "classification": "source-reported ceiling screen", "experiment_id": "x-race", "graded_n": 48, "planned_n": 48, "reported_passes": 48, "reps_per_arm": 2, "task_n": 6}, "round_id": "x-race", "round_page": "rounds/x-race/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-race-reported-x-race-selector-check", "claim_status": "INTERIM", "comparison": null, "denominator": 24, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 24, "population": "reported_denominator"}], "evidence": [{"path": "rounds/x-race/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-race/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported selector check; no public case rows", "experiment_id": "x-race-selector-check", "reported_denominator": 24}, "round_id": "x-race", "round_page": "rounds/x-race/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-race-reported-x-race-decision", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-race/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-race/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "development decision", "experiment_id": "x-race-decision", "reported_results": "no multi-agent arm promoted"}, "round_id": "x-race", "round_page": "rounds/x-race/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-parallel-reported-x-parallel-v03", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2, "population": "task_n"}], "evidence": [{"path": "rounds/x-parallel/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-parallel/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "solo vs planned DAG", "classification": "source-reported exploratory result", "experiment_id": "x-parallel-v03", "reported_results": "4/4 vs 4/4", "reps_per_arm": 2, "task_n": 2}, "round_id": "x-parallel", "round_page": "rounds/x-parallel/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-parallel-reported-x-parallel-v03-cost", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-parallel/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-parallel/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported summary", "experiment_id": "x-parallel-v03-cost", "reported_results": "about 1.4× wall; about 1.9× requests"}, "round_id": "x-parallel", "round_page": "rounds/x-parallel/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-parallel-reported-x-parallel-token-summary", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-parallel/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-parallel/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H30"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate source-reported token summary; raw rows unavailable", "experiment_id": "x-parallel-token-summary", "reported_results": "1.84× input tokens; 2.50× output tokens"}, "round_id": "x-parallel", "round_page": "rounds/x-parallel/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-minikogen-reported-x-minikogen-mechanism-pairs", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-minikogen/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-minikogen/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H31", "H32"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported small mechanism panel; arm map incomplete", "experiment_id": "x-minikogen-mechanism-pairs", "pair_count": 5}, "round_id": "x-minikogen", "round_page": "rounds/x-minikogen/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-minikogen-reported-x-minikogen-test-first-removal", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-minikogen/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-minikogen/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H31", "H32"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported; outcome and wall rows unavailable", "experiment_id": "x-minikogen-test-first-removal", "reported_results": "4/4 vs 4/4"}, "round_id": "x-minikogen", "round_page": "rounds/x-minikogen/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-minikogen-reported-x-minikogen-frozen-decision", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-minikogen/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-minikogen/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H29", "H31", "H32"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported small screen", "cost": "higher; amount not recovered", "experiment_id": "x-minikogen-frozen-decision", "reported_results": "strict acceptance 2/4 to 4/4"}, "round_id": "x-minikogen", "round_page": "rounds/x-minikogen/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-surface-reported-x-surface-feedback-format", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 16, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 8, "population": "completed_n"}], "evidence": [{"path": "rounds/x-surface/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-surface/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H31", "H33", "H110"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "partial cohort", "completed_n": 8, "experiment_id": "x-surface-feedback-format", "planned_n": 16}, "round_id": "x-surface", "round_page": "rounds/x-surface/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-surface-reported-x-surface-fixture-repair", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-surface/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-surface/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H31", "H33", "H110"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "small tied fixture subset", "experiment_id": "x-surface-fixture-repair", "reported_results": "2/2 vs 2/2 by model"}, "round_id": "x-surface", "round_page": "rounds/x-surface/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-surface-reported-x-surface-remaining-rep", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-surface/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-surface/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H31", "H33", "H110"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "incomplete", "experiment_id": "x-surface-remaining-rep", "reported_results": "unrun at cited cutoff"}, "round_id": "x-surface", "round_page": "rounds/x-surface/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-context-reported-x-context", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-context/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-context/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H51", "H54", "H55", "H56", "H57", "H58"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source summary only; exact cohort not recovered", "experiment_id": "x-context", "finished_n": null, "graded_n": null, "itt_n": null, "planned_n": null, "reported_status": "inconclusive", "started_n": null}, "round_id": "x-context", "round_page": "rounds/x-context/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-context-reported-x-context-admission-guard", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-context/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-context/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H51", "H54", "H55", "H56", "H57", "H58"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "candidate label; no effect estimate", "experiment_id": "x-context-admission-guard", "reported_status": "transfer candidate"}, "round_id": "x-context", "round_page": "rounds/x-context/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-compaction-reported-x-compaction-traces", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 60, "population": "run_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "per_arm_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "compacted_n"}], "evidence": [{"path": "rounds/x-compaction/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-compaction/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H54"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arms": "kh; Codex; Pi", "classification": "source-reported trace census", "compacted_n": 0, "experiment_id": "x-compaction-traces", "per_arm_n": 20, "run_n": 60}, "round_id": "x-compaction", "round_page": "rounds/x-compaction/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-compaction-reported-x-compaction-peak-input", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-compaction/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-compaction/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H54"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported medians; raw traces unavailable", "experiment_id": "x-compaction-peak-input"}, "round_id": "x-compaction", "round_page": "rounds/x-compaction/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-compaction-reported-x-compaction-elision-review", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-compaction/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-compaction/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H54"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "preliminary source-reported review", "experiment_id": "x-compaction-elision-review", "reported_results": "14/18 facts retained"}, "round_id": "x-compaction", "round_page": "rounds/x-compaction/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-compaction-reported-e06-context-pilot-link", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-compaction/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/x-compaction/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H54"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate study; see e06-context", "experiment_id": "e06-context-pilot-link", "reported_results": "1 pilot; no adjacent comparison"}, "round_id": "x-compaction", "round_page": "rounds/x-compaction/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-continuation-reported-x-continuation-live", "claim_status": "NOT-RUN", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-continuation/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-continuation/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H31", "H65"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "blocked before model execution", "experiment_id": "x-continuation-live", "real_model_attempts": 0}, "round_id": "x-continuation", "round_page": "rounds/x-continuation/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-continuation-reported-x-continuation-scripted-provider", "claim_status": "NOT-RUN", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-continuation/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-continuation/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H31", "H65"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "offline mechanical test; not a live task outcome", "experiment_id": "x-continuation-scripted-provider", "reported_results": "3/6 before; 6/6 after"}, "round_id": "x-continuation", "round_page": "rounds/x-continuation/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-recovery-reported-x-recovery", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/x-recovery/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-recovery/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H65"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "unverified", "experiment_id": "x-recovery", "finished_n": null, "graded_n": null, "itt_n": null, "planned_n": null, "reported_status": "ledger identified; per-arm extraction required", "started_n": null}, "round_id": "x-recovery", "round_page": "rounds/x-recovery/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-mining-reported-x-mining-observation-inventory", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2608, "population": "observations_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 2152, "population": "pass_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 295, "population": "fail_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 54, "population": "infra_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 107, "population": "pending_n"}], "evidence": [{"path": "rounds/x-mining/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/x-mining/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H80", "H84", "H89"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "source-reported mixed-version inventory", "experiment_id": "x-mining-observation-inventory", "fail_n": 295, "infra_n": 54, "observations_n": 2608, "pass_n": 2152, "pending_n": 107}, "round_id": "x-mining", "round_page": "rounds/x-mining/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-mining-reported-x-mining-usage-audit", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2649, "population": "observations_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 124, "population": "missing_usage_n"}], "evidence": [{"path": "rounds/x-mining/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/x-mining/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H80", "H84", "H89"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate source-reported usage denominator", "experiment_id": "x-mining-usage-audit", "missing_usage_n": 124, "observations_n": 2649}, "round_id": "x-mining", "round_page": "rounds/x-mining/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-x-mining-reported-x-mining-x09", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 14, "population": "completed_n"}], "evidence": [{"path": "rounds/x-mining/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/x-mining/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H80", "H84", "H89"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate diagnostic screen", "completed_n": 14, "experiment_id": "x-mining-x09", "reported_status": "inconclusive"}, "round_id": "x-mining", "round_page": "rounds/x-mining/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-e06-context-reported-e06-context-pilot", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 1, "population": "pilot_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "adjacent_comparison_n"}], "evidence": [{"path": "rounds/e06-context/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/e06-context/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H54", "H55"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"adjacent_comparison_n": 0, "classification": "source-reported single pilot", "experiment_id": "e06-context-pilot", "pilot_n": 1, "reported_effective_compactions": 6}, "round_id": "e06-context", "round_page": "rounds/e06-context/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-ctx-store-reported-ctx-store-ranked-paths", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 5, "population": "store_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 8, "population": "query_n"}], "evidence": [{"path": "rounds/ctx-store/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/ctx-store/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H148", "H149", "H150"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "offline source-reported study", "experiment_id": "ctx-store-ranked-paths", "query_n": 8, "reported_result": "all stores tied on ranked paths", "store_n": 5}, "round_id": "ctx-store", "round_page": "rounds/ctx-store/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-ctx-store-reported-ctx-store-recall", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/ctx-store/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/ctx-store/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H148", "H149", "H150"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "offline source-reported metric", "experiment_id": "ctx-store-recall"}, "round_id": "ctx-store", "round_page": "rounds/ctx-store/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-ctx-store-reported-ctx-store-freshness", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 4266, "population": "record_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 25, "population": "checks_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 25, "population": "checks_passed_n"}], "evidence": [{"path": "rounds/ctx-store/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/ctx-store/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H148", "H149", "H150"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"checks_n": 25, "checks_passed_n": 25, "classification": "separate freshness audit; not Build success", "experiment_id": "ctx-store-freshness", "record_n": 4266}, "round_id": "ctx-store", "round_page": "rounds/ctx-store/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-e09-tidewave-reported-e09-locator-gates", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 13, "population": "locator_variant_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 49, "population": "development_case_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 70, "population": "source_case_set_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 7, "population": "stack_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "passed_variants_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "model_dispatches_n"}], "evidence": [{"path": "rounds/e09-tidewave/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/e09-tidewave/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H55", "H60"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "offline source-reported diagnostic", "development_case_n": 49, "experiment_id": "e09-locator-gates", "gate_precision": "at least 75%", "gate_recall": "100%", "locator_variant_n": 13, "model_dispatches_n": 0, "passed_variants_n": 0, "source_case_set_n": 70, "stack_n": 7}, "round_id": "e09-tidewave", "round_page": "rounds/e09-tidewave/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-e09-tidewave-reported-e09-tidewave-runtime", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 12, "population": "planned_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "stopped_after_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "unprompted_calls_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 17, "population": "instructed_calls_n"}], "evidence": [{"path": "rounds/e09-tidewave/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/e09-tidewave/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H55", "H60"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate partial runtime-tool screen", "experiment_id": "e09-tidewave-runtime", "instructed_calls_n": 17, "planned_n": 12, "stopped_after_n": 6, "unprompted_calls_n": 0}, "round_id": "e09-tidewave", "round_page": "rounds/e09-tidewave/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-overall", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 4, "population": "host_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 30, "population": "invalid_grade_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 7, "population": "invalid_task_n"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "syn-33_invalid_n"}], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/0"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "approximate cross-lane inventory; invalid grades excluded from model-loss claims", "experiment_id": "night-overall", "host_n": 4, "invalid_grade_n": 30, "invalid_task_n": 7, "syn-33_invalid_n": 20}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-p1", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/1"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "keep separate from P2", "experiment_id": "night-p1", "planned_n": null, "reported_results": "not recovered"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-p2", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 52, "population": "admitted_task_n"}], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/2"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"admitted_task_n": 52, "classification": "valid-only source-reported summary; exact cells unavailable", "experiment_id": "night-p2"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-p3a", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/3"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate exploratory lane", "experiment_id": "night-p3a", "reported_results": "relay 5/5; L1 3/5"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-p3c", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/4"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "separate exploratory lane; do not pool with P3a", "experiment_id": "night-p3c", "reported_results": "relay 4/5; L1 5/5"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-p5", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/5"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "selected subset; exact task mapping unavailable", "experiment_id": "night-p5", "reported_results": "relay 2/9; L1 0/9 on selected failures; seven Rails tasks reportedly all failed"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-parallel-read", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/6"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "lane mapping and raw request counts unavailable", "experiment_id": "night-parallel-read", "reported_results": "5/5 tie; about 20–30% fewer requests"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Retain each experiment record as a separate observation; do not promote offline, pilot, partial, or mixed-version counts to Build efficacy.", "arm": null, "claim_id": "CL-WA-night-2026-10-01-reported-night-pi-mining", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"path": "rounds/night-2026-10-01/reported-data.json", "pointer": "/experiment_records/7"}, {"path": "rounds/night-2026-10-01/README.md", "selector": "reproduction and limitation statement"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H19", "H21", "H29", "H31", "H80", "H81", "H82", "H84", "H85", "H86", "H87", "H88", "H89", "H90"], "interpretation": "No exact cell/task ID join was recovered in the public result exports.", "metric": "documentary_experiment_record", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"classification": "not-run model arm", "experiment_id": "night-pi-mining", "reported_results": "Pi 1.0.0 installed for mining, not run as a benchmark arm"}, "round_id": "night-2026-10-01", "round_page": "rounds/night-2026-10-01/README.md", "schema_version": "1.0", "unit": "source-reported analysis units"}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-hc-1-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 40, "population": "graded scored/guard observations"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "ladder arm"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "short ladder arm"}], "evidence": [{"path": "rounds/hc-1/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/hc-1/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H75", "H136"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "hill_climb_lifecycle_and_outcome", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"graded_n": 40, "ladder_denominator": 20, "ladder_passes": 19, "short_ladder_denominator": 20, "short_ladder_passes": 17}, "round_id": "hc-1", "round_page": "rounds/hc-1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-hc-2-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 60, "population": "v2 decision cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 30, "population": "v1 guard cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 90, "population": "all scored and guard rows"}], "evidence": [{"path": "rounds/hc-2/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/hc-2/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H75", "H136"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "hill_climb_lifecycle_and_outcome", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"diverse_denominator": 20, "diverse_passes": 13, "floor_denominator": 20, "floor_passes": 13, "graded_n": 90, "guard_cells": 30, "guard_diverse_denominator": 10, "guard_diverse_passes": 10, "guard_floor_denominator": 10, "guard_floor_passes": 7, "guard_ladder_denominator": 10, "guard_ladder_passes": 9, "ladder_denominator": 20, "ladder_passes": 16, "v2_scored_cells": 60}, "round_id": "hc-2", "round_page": "rounds/hc-2/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-hc-3-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 40, "population": "v2 scored decision rows"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "guard cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 60, "population": "all rows"}], "evidence": [{"path": "rounds/hc-3/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/hc-3/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H75", "H136"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "hill_climb_lifecycle_and_outcome", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"candidate_reranks": 0, "dropped_tests": 147, "edge_test_denominator": 20, "edge_test_passes": 13, "generated_tests": 197, "graded_n": 60, "guard_cells": 20, "kept_tests": 50, "ladder_denominator": 20, "ladder_passes": 16, "selected_repairs": 0, "triggered_repairs": 4, "v2_scored_cells": 40}, "round_id": "hc-3", "round_page": "rounds/hc-3/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-hc-4-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 29, "population": "result directories"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 43, "population": "expected retained slots"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 9, "population": "started diagnostics inside retained slots"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "stopped cells"}], "evidence": [{"path": "rounds/hc-4/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/hc-4/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H75", "H136", "H138"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "hill_climb_lifecycle_and_outcome", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"owner_stopped_studio_cells": 6, "result_directories": 29, "retained_expected_n": 43, "started_diagnostics_in_retained_n": 9}, "round_id": "hc-4", "round_page": "rounds/hc-4/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-r72-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 456, "population": "source-reported scored specs headline"}, {"evidence_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export", "n": 84, "population": "P1 Kogen scored cells", "source_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export"}, {"evidence_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export", "n": 84, "population": "P1 Codex scored cells", "source_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export"}, {"evidence_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export", "n": 13, "population": "P3 Kogen partial cells", "source_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export"}, {"evidence_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export", "n": 11, "population": "P3 Codex partial cells", "source_status": "reproducible from public L0 reconciliation CSV; not independently joined to canonical official outcome export"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "P2 scored cells"}], "evidence": [{"path": "rounds/r72/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/r72/MEASURED.md", "selector": "measurement and denominator definition"}, {"path": "rounds/l0-reconcile/data/r72-p1.csv", "selector": "all 168 recovered P1 rows; public reconciliation input"}, {"path": "rounds/l0-reconcile/data/r72-p3-partial.csv", "selector": "all 24 recovered partial P3 rows"}, {"path": "reproduce/reconcile_l0.py", "selector": "r72()"}], "evidence_status": "Reproducible from public L0 reconciliation CSVs for R72 P1 (64/84 vs 60/84) and partial P3 (12/13 vs 10/11); not independently joined to the canonical official outcome export (0 matching scored rows). Initial design and complete lifecycle remain source-reported, not reproducible from public data.", "exact_cell_ids": [], "hypothesis_ids": ["H113", "H136", "H138"], "interpretation": "L0 publishes cell-level CSV inputs that reproduce P1 and partial P3 tallies. The official outcome export still has no matching scored rows, and the complete planned/start/finish/ITT partition is unrecovered. INTERIM and descriptive only; no superiority claim.", "metric": "early_stop_round_denominators", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reconciliation_status": "Public L0 CSV analysis is reproducible; official outcome join status remains explicitly scoped in evidence_status.", "reported_values": {"headline_scored_specs_n": 456, "official_smokes_denominator": 6, "official_smokes_passed": 6, "p1_codex_denominator": 84, "p1_codex_passes": 60, "p1_kogen_denominator": 84, "p1_kogen_passes": 64, "p2_scored_n": 0, "p3_codex_denominator": 11, "p3_codex_passes": 10, "p3_kogen_denominator": 13, "p3_kogen_passes": 12}, "round_id": "r72", "round_page": "rounds/r72/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-frontier-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 70, "population": "prepared design cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 60, "population": "unstarted cells trimmed"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "released in public snapshot"}], "evidence": [{"path": "rounds/frontier/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/frontier/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H90", "H138", "H151"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "frontier_lane_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"active_storage_closure": "unverified", "duplicate_trim_unstarted_n": 60, "prepared_cells_n": 70, "released_cells_n": 0}, "round_id": "frontier", "round_page": "rounds/frontier/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-studio-kogen-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 4, "population": "smoke attempts"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "bulk scored cells"}], "evidence": [{"path": "rounds/studio-kogen/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/studio-kogen/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H65", "H87", "H89"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "smoke_only_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"bulk_cells_n": 0, "provided_retry_denominator": 1, "provided_retry_passes": 1, "shaped_retry_invalid_n": 1, "smoke_attempts_n": 4}, "round_id": "studio-kogen", "round_page": "rounds/studio-kogen/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-studio-kogen-80a4-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "smoke attempts"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "bulk scored cells"}], "evidence": [{"path": "rounds/studio-kogen-80a4/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/studio-kogen-80a4/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H65", "H87", "H89"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "smoke_only_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"a1_smoke_denominator": 3, "a1_smoke_passes": 3, "bulk_cells_n": 0, "smoke_attempts_n": 6}, "round_id": "studio-kogen-80a4", "round_page": "rounds/studio-kogen-80a4/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-studio-shaper-pilot-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2, "population": "smoke attempts"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 24, "population": "planned scored cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "released scored cells"}], "evidence": [{"path": "rounds/studio-shaper-pilot/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/studio-shaper-pilot/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H65", "H87", "H88"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "smoke_only_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"scored_planned_n": 24, "scored_released_n": 0, "smoke_invalids": 1, "smoke_passes": 1}, "round_id": "studio-shaper-pilot", "round_page": "rounds/studio-shaper-pilot/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-r57-studio-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2, "population": "invalid smokes"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 34, "population": "planned catalogue cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "released catalogue cells"}], "evidence": [{"path": "rounds/r57-studio/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/r57-studio/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H87", "H88", "H89"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "setup_invalid_pilot", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"catalogue_planned_n": 34, "catalogue_released_n": 0, "setup_invalid_smokes_n": 2}, "round_id": "r57-studio", "round_page": "rounds/r57-studio/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-r57-t90-studio-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 2, "population": "smokes"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 34, "population": "planned scored cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "released scored cells"}], "evidence": [{"path": "rounds/r57-t90-studio/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/r57-t90-studio/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H87", "H88", "H89"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "smoke_only_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"archive_smoke_passes": 1, "broadcast_smoke_invalids": 1, "scored_planned_n": 34, "scored_released_n": 0, "smokes_n": 2}, "round_id": "r57-t90-studio", "round_page": "rounds/r57-t90-studio/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep scored cells, smokes, guards, prepared slots, started diagnostics, stopped slots, and result-directory counts separate.", "arm": null, "claim_id": "CL-WA-r58x-studio-lifecycle-and-outcome", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "early checkpoint released cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 36, "population": "later finished scored rows"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 18, "population": "rows per arm"}], "evidence": [{"path": "rounds/r58x-studio/README.md", "selector": "reported lifecycle and status"}, {"path": "rounds/r58x-studio/MEASURED.md", "selector": "measurement and denominator definition"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H77", "H87", "H88", "H138"], "interpretation": "Exact per-cell outcome joins are unavailable; these observations do not establish a general product result.", "metric": "partial_round_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"arm_count": 2, "early_checkpoint_released_n": 0, "final_arm_outcomes": "unreconciled", "finished_per_arm": 18, "later_finished_scored_n": 36, "smokes_n": 2}, "round_id": "r58x-studio", "round_page": "rounds/r58x-studio/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-r70-original-rve-elixir-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 10, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 10, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-original-rve", "scored": true, "stack": "elixir"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-elixir__r12", "codex__gpt-6-luna__max__default__r70-1-elixir__r15", "codex__gpt-6-luna__max__default__r70-1-elixir__r6", "codex__gpt-6-luna__max__default__r70-3-elixir__r14", "codex__gpt-6-luna__max__default__r70-3-elixir__r7", "codex__gpt-6-luna__max__default__r70-4-elixir__r10", "codex__gpt-6-luna__max__default__r70-4-elixir__r13", "codex__gpt-6-luna__max__default__r70-4-elixir__r9", "codex__gpt-6-luna__max__default__r70-6-elixir__r11", "codex__gpt-6-luna__max__default__r70-6-elixir__r8"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 2, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-original-rve", "full_passes": 2, "graded_scored_cells": 10, "stack": "elixir"}, "round_id": "r70-rve", "round_page": "rounds/r70-rve/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-r70-original-rve-go-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 10, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 10, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-original-rve", "scored": true, "stack": "go"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-go__r16", "codex__gpt-6-luna__max__default__r70-1-go__r22", "codex__gpt-6-luna__max__default__r70-1-go__r25", "codex__gpt-6-luna__max__default__r70-3-go__r17", "codex__gpt-6-luna__max__default__r70-3-go__r24", "codex__gpt-6-luna__max__default__r70-4-go__r19", "codex__gpt-6-luna__max__default__r70-4-go__r20", "codex__gpt-6-luna__max__default__r70-4-go__r23", "codex__gpt-6-luna__max__default__r70-6-go__r18", "codex__gpt-6-luna__max__default__r70-6-go__r21"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 7, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-original-rve", "full_passes": 7, "graded_scored_cells": 10, "stack": "go"}, "round_id": "r70-rve", "round_page": "rounds/r70-rve/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-r70-original-rve-rust-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 10, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 10, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-original-rve", "scored": true, "stack": "rust"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-rust__r12", "codex__gpt-6-luna__max__default__r70-1-rust__r15", "codex__gpt-6-luna__max__default__r70-1-rust__r6", "codex__gpt-6-luna__max__default__r70-3-rust__r14", "codex__gpt-6-luna__max__default__r70-3-rust__r7", "codex__gpt-6-luna__max__default__r70-4-rust__r10", "codex__gpt-6-luna__max__default__r70-4-rust__r13", "codex__gpt-6-luna__max__default__r70-4-rust__r9", "codex__gpt-6-luna__max__default__r70-6-rust__r11", "codex__gpt-6-luna__max__default__r70-6-rust__r8"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 8, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-original-rve", "full_passes": 8, "graded_scored_cells": 10, "stack": "rust"}, "round_id": "r70-rve", "round_page": "rounds/r70-rve/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-r70-original-rve-ts-bun-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 10, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 10, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-original-rve", "scored": true, "stack": "ts-bun"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-ts-bun__r16", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r22", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r25", "codex__gpt-6-luna__max__default__r70-3-ts-bun__r17", "codex__gpt-6-luna__max__default__r70-3-ts-bun__r24", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r19", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r20", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r23", "codex__gpt-6-luna__max__default__r70-6-ts-bun__r18", "codex__gpt-6-luna__max__default__r70-6-ts-bun__r21"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 7, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-original-rve", "full_passes": 7, "graded_scored_cells": 10, "stack": "ts-bun"}, "round_id": "r70-rve", "round_page": "rounds/r70-rve/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-rerun-r70-rve-rerun-elixir-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 6, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 6, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-rerun", "scored": true, "stack": "elixir"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-elixir-fe2__r12", "codex__gpt-6-luna__max__default__r70-1-elixir-fe2__r15", "codex__gpt-6-luna__max__default__r70-1-elixir-fe2__r6", "codex__gpt-6-luna__max__default__r70-4-elixir-fe2__r10", "codex__gpt-6-luna__max__default__r70-4-elixir-fe2__r13", "codex__gpt-6-luna__max__default__r70-4-elixir-fe2__r9"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 3, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-rerun", "full_passes": 3, "graded_scored_cells": 6, "stack": "elixir"}, "round_id": "r70-rve-rerun", "round_page": "rounds/r70-rve-rerun/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-rerun-r70-rve-rerun-go-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 3, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 3, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-rerun", "scored": true, "stack": "go"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-4-go-fe2__r10", "codex__gpt-6-luna__max__default__r70-4-go-fe2__r13", "codex__gpt-6-luna__max__default__r70-4-go-fe2__r9"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 3, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-rerun", "full_passes": 3, "graded_scored_cells": 3, "stack": "go"}, "round_id": "r70-rve-rerun", "round_page": "rounds/r70-rve-rerun/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-rerun-r70-rve-rerun-rust-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 3, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 3, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-rerun", "scored": true, "stack": "rust"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-4-rust-fe2__r10", "codex__gpt-6-luna__max__default__r70-4-rust-fe2__r13", "codex__gpt-6-luna__max__default__r70-4-rust-fe2__r9"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 3, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-rerun", "full_passes": 3, "graded_scored_cells": 3, "stack": "rust"}, "round_id": "r70-rve-rerun", "round_page": "rounds/r70-rve-rerun/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-rerun-r70-rve-rerun-ts-bun-full-pass", "claim_status": "CONFOUNDED", "comparison": null, "denominator": 3, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 3, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-rerun", "scored": true, "stack": "ts-bun"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-4-ts-bun-fe2__r10", "codex__gpt-6-luna__max__default__r70-4-ts-bun-fe2__r13", "codex__gpt-6-luna__max__default__r70-4-ts-bun-fe2__r9"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 2, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-rerun", "full_passes": 2, "graded_scored_cells": 3, "stack": "ts-bun"}, "round_id": "r70-rve-rerun", "round_page": "rounds/r70-rve-rerun/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-ext-r70-rve-ext-elixir-full-pass", "claim_status": "VALID", "comparison": null, "denominator": 8, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 8, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-ext", "scored": true, "stack": "elixir"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-2-elixir__r31", "codex__gpt-6-luna__max__default__r70-2-elixir__r32", "codex__gpt-6-luna__max__default__r70-2-elixir__r33", "codex__gpt-6-luna__max__default__r70-5-elixir__r31", "codex__gpt-6-luna__max__default__r70-5-elixir__r32", "codex__gpt-6-luna__max__default__r70-7-elixir__r31", "codex__gpt-6-luna__max__default__r70-7-elixir__r32", "codex__gpt-6-luna__max__default__r70-7-elixir__r33"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 5, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-ext", "full_passes": 5, "graded_scored_cells": 8, "stack": "elixir"}, "round_id": "r70-rve-ext", "round_page": "rounds/r70-rve-ext/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-ext-r70-rve-ext-go-full-pass", "claim_status": "VALID", "comparison": null, "denominator": 8, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 8, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-ext", "scored": true, "stack": "go"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-2-go__r31", "codex__gpt-6-luna__max__default__r70-2-go__r32", "codex__gpt-6-luna__max__default__r70-2-go__r33", "codex__gpt-6-luna__max__default__r70-5-go__r31", "codex__gpt-6-luna__max__default__r70-5-go__r32", "codex__gpt-6-luna__max__default__r70-5-go__r33", "codex__gpt-6-luna__max__default__r70-7-go__r31", "codex__gpt-6-luna__max__default__r70-7-go__r32"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 6, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-ext", "full_passes": 6, "graded_scored_cells": 8, "stack": "go"}, "round_id": "r70-rve-ext", "round_page": "rounds/r70-rve-ext/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-ext-r70-rve-ext-rust-full-pass", "claim_status": "VALID", "comparison": null, "denominator": 7, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 7, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-ext", "scored": true, "stack": "rust"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-2-rust__r31", "codex__gpt-6-luna__max__default__r70-2-rust__r32", "codex__gpt-6-luna__max__default__r70-5-rust__r31", "codex__gpt-6-luna__max__default__r70-5-rust__r32", "codex__gpt-6-luna__max__default__r70-7-rust__r31", "codex__gpt-6-luna__max__default__r70-7-rust__r32", "codex__gpt-6-luna__max__default__r70-7-rust__r33"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 6, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-ext", "full_passes": 6, "graded_scored_cells": 7, "stack": "rust"}, "round_id": "r70-rve-ext", "round_page": "rounds/r70-rve-ext/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Count exact scored cell IDs in the named cohort and stack; full pass is the official outcome value pass; do not pool original, rerun, and extension cohorts.", "arm": null, "claim_id": "CL-WA-r70-rve-ext-r70-rve-ext-ts-bun-full-pass", "claim_status": "VALID", "comparison": null, "denominator": 6, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 6, "population": "scored official cells in supplemental test-count ledger"}], "evidence": [{"filter": {"cohort": "r70-rve-ext", "scored": true, "stack": "ts-bun"}, "path": "results/test-counts.jsonl"}, {"path": "results/README.md", "selector": "supplemental Round 70 receipts"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-2-ts-bun__r31", "codex__gpt-6-luna__max__default__r70-2-ts-bun__r32", "codex__gpt-6-luna__max__default__r70-5-ts-bun__r31", "codex__gpt-6-luna__max__default__r70-5-ts-bun__r32", "codex__gpt-6-luna__max__default__r70-7-ts-bun__r31", "codex__gpt-6-luna__max__default__r70-7-ts-bun__r32"], "hypothesis_ids": ["H114"], "interpretation": "The result is limited to this cohort. Original/RvE and rerun rows retain documented parity confounds; extension is descriptive.", "metric": "official_full_pass", "numerator": 6, "publication_status": "STATUS_ONLY", "reported_values": {"cohort": "r70-rve-ext", "full_passes": 6, "graded_scored_cells": 6, "stack": "ts-bun"}, "round_id": "r70-rve-ext", "round_page": "rounds/r70-rve-ext/README.md", "schema_version": "1.0", "unit": "scored official cells"}
{"analysis_rule": "Do not assign IDs or outcomes to the five unjoined deliveries; keep the scored export cohort and reported ITT extension distinct.", "arm": null, "claim_id": "CL-WA-r70-rve-ungraded-delivery-itt-gap", "claim_status": "CONFOUNDED", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public data", "n": 40, "population": "public test-count scored IDs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 5, "population": "additional ungraded deliveries without public cell IDs"}], "evidence": [{"filter": {"cohort": "r70-original-rve", "scored": true}, "path": "results/test-counts.jsonl"}, {"path": "rounds/r70-rve/README.md", "selector": "Cohort accounting"}], "evidence_status": "reproducible from public data for 40 scored cells; source-reported, not reproducible from public data for the five extra deliveries", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-1-elixir__r12", "codex__gpt-6-luna__max__default__r70-1-elixir__r15", "codex__gpt-6-luna__max__default__r70-1-elixir__r6", "codex__gpt-6-luna__max__default__r70-1-go__r16", "codex__gpt-6-luna__max__default__r70-1-go__r22", "codex__gpt-6-luna__max__default__r70-1-go__r25", "codex__gpt-6-luna__max__default__r70-1-rust__r12", "codex__gpt-6-luna__max__default__r70-1-rust__r15", "codex__gpt-6-luna__max__default__r70-1-rust__r6", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r16", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r22", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r25", "codex__gpt-6-luna__max__default__r70-3-elixir__r14", "codex__gpt-6-luna__max__default__r70-3-elixir__r7", "codex__gpt-6-luna__max__default__r70-3-go__r17", "codex__gpt-6-luna__max__default__r70-3-go__r24", "codex__gpt-6-luna__max__default__r70-3-rust__r14", "codex__gpt-6-luna__max__default__r70-3-rust__r7", "codex__gpt-6-luna__max__default__r70-3-ts-bun__r17", "codex__gpt-6-luna__max__default__r70-3-ts-bun__r24", "codex__gpt-6-luna__max__default__r70-4-elixir__r10", "codex__gpt-6-luna__max__default__r70-4-elixir__r13", "codex__gpt-6-luna__max__default__r70-4-elixir__r9", "codex__gpt-6-luna__max__default__r70-4-go__r19", "codex__gpt-6-luna__max__default__r70-4-go__r20", "codex__gpt-6-luna__max__default__r70-4-go__r23", "codex__gpt-6-luna__max__default__r70-4-rust__r10", "codex__gpt-6-luna__max__default__r70-4-rust__r13", "codex__gpt-6-luna__max__default__r70-4-rust__r9", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r19", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r20", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r23", "codex__gpt-6-luna__max__default__r70-6-elixir__r11", "codex__gpt-6-luna__max__default__r70-6-elixir__r8", "codex__gpt-6-luna__max__default__r70-6-go__r18", "codex__gpt-6-luna__max__default__r70-6-go__r21", "codex__gpt-6-luna__max__default__r70-6-rust__r11", "codex__gpt-6-luna__max__default__r70-6-rust__r8", "codex__gpt-6-luna__max__default__r70-6-ts-bun__r18", "codex__gpt-6-luna__max__default__r70-6-ts-bun__r21"], "hypothesis_ids": ["H114"], "interpretation": "The complete ITT denominator is unresolved.", "metric": "unmatched_ungraded_delivery_count", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"additional_ungraded_deliveries_without_joinable_ids": 5, "itt_status": "unresolved", "reconciled_scored_cell_ids": 40}, "round_id": "r70-rve", "round_page": "rounds/r70-rve/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "The test fraction is a smoke diagnostic; scored is false and it is excluded from efficacy denominators.", "arm": null, "claim_id": "CL-WA-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-go-r90", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": 25, "denominator_variants": [], "evidence": [{"filter": {"cell_id": "codex__gpt-6-luna__max__default__r70-8-go__r90", "cohort": "r70-task8-v1-smoke"}, "path": "results/test-counts.jsonl"}, {"filter": {"cohort": "r70-task8-v1"}, "path": "results/controls.jsonl"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-8-go__r90"], "hypothesis_ids": ["H114"], "interpretation": "Descriptive unscored smoke only. No task-8 language or prompt efficacy claim follows from an unscored smoke.", "metric": "smoke_test_fraction", "numerator": 1, "publication_status": "STATUS_ONLY", "reported_values": {"official_smoke_outcome": "fail", "scored": false, "stack": "go", "tests_passed": 1, "tests_total": 25}, "round_id": "r70-task8", "round_page": "rounds/r70-task8/README.md", "schema_version": "1.0", "unit": "hidden-test assertions in unscored smoke"}
{"analysis_rule": "The test fraction is a smoke diagnostic; scored is false and it is excluded from efficacy denominators.", "arm": null, "claim_id": "CL-WA-r70-task8v1-smoke-codex-gpt-6-luna-max-default-r70-8-rust-r90", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": 25, "denominator_variants": [], "evidence": [{"filter": {"cell_id": "codex__gpt-6-luna__max__default__r70-8-rust__r90", "cohort": "r70-task8-v1-smoke"}, "path": "results/test-counts.jsonl"}, {"filter": {"cohort": "r70-task8-v1"}, "path": "results/controls.jsonl"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-8-rust__r90"], "hypothesis_ids": ["H114"], "interpretation": "Descriptive unscored smoke only. No task-8 language or prompt efficacy claim follows from an unscored smoke.", "metric": "smoke_test_fraction", "numerator": 1, "publication_status": "STATUS_ONLY", "reported_values": {"official_smoke_outcome": "fail", "scored": false, "stack": "rust", "tests_passed": 1, "tests_total": 25}, "round_id": "r70-task8", "round_page": "rounds/r70-task8/README.md", "schema_version": "1.0", "unit": "hidden-test assertions in unscored smoke"}
{"analysis_rule": "The test fraction is a smoke diagnostic; scored is false and it is excluded from efficacy denominators.", "arm": null, "claim_id": "CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-go-v2-r1", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": 25, "denominator_variants": [], "evidence": [{"filter": {"cell_id": "codex__gpt-6-luna__max__default__r70-8-go-v2__r1", "cohort": "r70-task8-v2-smoke"}, "path": "results/test-counts.jsonl"}, {"filter": {"cohort": "r70-task8-v2"}, "path": "results/controls.jsonl"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-8-go-v2__r1"], "hypothesis_ids": ["H114"], "interpretation": "Descriptive unscored smoke only. No task-8 language or prompt efficacy claim follows from an unscored smoke.", "metric": "smoke_test_fraction", "numerator": 1, "publication_status": "STATUS_ONLY", "reported_values": {"official_smoke_outcome": "fail", "scored": false, "stack": "go", "tests_passed": 1, "tests_total": 25}, "round_id": "r70-task8", "round_page": "rounds/r70-task8/README.md", "schema_version": "1.0", "unit": "hidden-test assertions in unscored smoke"}
{"analysis_rule": "The test fraction is a smoke diagnostic; scored is false and it is excluded from efficacy denominators.", "arm": null, "claim_id": "CL-WA-r70-task8v2-smoke-codex-gpt-6-luna-max-default-r70-8-rust-v2-r1", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": 25, "denominator_variants": [], "evidence": [{"filter": {"cell_id": "codex__gpt-6-luna__max__default__r70-8-rust-v2__r1", "cohort": "r70-task8-v2-smoke"}, "path": "results/test-counts.jsonl"}, {"filter": {"cohort": "r70-task8-v2"}, "path": "results/controls.jsonl"}], "evidence_status": "reproducible from public data", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-8-rust-v2__r1"], "hypothesis_ids": ["H114"], "interpretation": "Descriptive unscored smoke only. No task-8 language or prompt efficacy claim follows from an unscored smoke.", "metric": "smoke_test_fraction", "numerator": 1, "publication_status": "STATUS_ONLY", "reported_values": {"official_smoke_outcome": "fail", "scored": false, "stack": "rust", "tests_passed": 1, "tests_total": 25}, "round_id": "r70-task8", "round_page": "rounds/r70-task8/README.md", "schema_version": "1.0", "unit": "hidden-test assertions in unscored smoke"}
{"analysis_rule": "Formal cells and preparation smokes are separate; zero formal observations define no outcome rate.", "arm": null, "claim_id": "CL-WA-campfire-lane1-formal-cell-lifecycle", "claim_status": "WITHDRAWN", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 30, "population": "planned formal cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "started formal cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "preparation smoke checks"}], "evidence": [{"path": "rounds/campfire-lane1/README.md", "selector": "design and lifecycle"}, {"path": "rounds/campfire-lane1/DESIGN.md", "selector": "registered design"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "WITHDRAWN before formal execution; no comparative runtime result is available.", "metric": "withdrawn_formal_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"finished_formal_cells": 0, "graded_formal_cells": 0, "itt_observations": 0, "planned_formal_cells": 30, "preparation_smokes": 0, "started_formal_cells": 0}, "round_id": "campfire-lane1", "round_page": "rounds/campfire-lane1/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Formal cells and preparation smokes are separate; zero formal observations define no outcome rate.", "arm": null, "claim_id": "CL-WA-campfire-lane2-formal-cell-lifecycle", "claim_status": "WITHDRAWN", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 50, "population": "planned formal cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "started formal cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 2, "population": "preparation smoke checks"}], "evidence": [{"path": "rounds/campfire-lane2/README.md", "selector": "design and lifecycle"}, {"path": "rounds/campfire-lane2/DESIGN.md", "selector": "registered design"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "WITHDRAWN before formal execution; no comparative runtime result is available.", "metric": "withdrawn_formal_lifecycle", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"finished_formal_cells": 0, "graded_formal_cells": 0, "itt_observations": 0, "planned_formal_cells": 50, "preparation_smokes": 2, "started_formal_cells": 0}, "round_id": "campfire-lane2", "round_page": "rounds/campfire-lane2/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep schedule capacity, started requests, terminal requests, uncensored completions, and censored/unstarted requests separate; quantiles use completed requests only.", "arm": null, "claim_id": "CL-WA-llm-latency-short-request-counts-and-endpoints", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public raw logs or measurement definition", "n": 36, "population": "scheduled_capacity_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 77, "population": "request_attempts_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 77, "population": "terminal_records_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 76, "population": "uncensored_completed_n"}], "evidence": [{"path": "rounds/llm-latency-short/raw/eu-20261005T223815Z.jsonl", "selector": "event rows; numeric client-visible timing only"}, {"path": "rounds/llm-latency-short/raw/us-20261005T223819Z.jsonl", "selector": "event rows; numeric client-visible timing only"}, {"path": "rounds/llm-latency-short/reproduce.py", "selector": "summary calculation"}, {"path": "rounds/llm-latency-short/MEASURED.md", "selector": "scheduled and observed populations"}], "evidence_status": "reproducible from public raw logs", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Client-visible timing is not provider transport latency, task success, or a Kogen Build efficacy result.", "metric": "client_visible_latency", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"censored_n": 1, "graded_n": 0, "per_arm": {"gpt-6-luna/high": {"attempts": 14, "complete": 13, "first_stdout_p50_s": 0.708, "first_stdout_p90_s": 0.987, "wall_p50_s": 5.331, "wall_p90_s": 6.722}, "gpt-6-luna/max": {"attempts": 12, "complete": 12, "first_stdout_p50_s": 0.67, "first_stdout_p90_s": 1.021, "wall_p50_s": 7.116, "wall_p90_s": 8.266}, "gpt-6-luna/medium": {"attempts": 15, "complete": 15, "first_stdout_p50_s": 0.723, "first_stdout_p90_s": 1.363, "wall_p50_s": 5.416, "wall_p90_s": 7.256}, "gpt-6-luna/xhigh": {"attempts": 12, "complete": 12, "first_stdout_p50_s": 0.653, "first_stdout_p90_s": 0.939, "wall_p50_s": 6.82, "wall_p90_s": 8.216}, "gpt-6.1-sol/high": {"attempts": 12, "complete": 12, "first_stdout_p50_s": 0.738, "first_stdout_p90_s": 1.002, "wall_p50_s": 7.342, "wall_p90_s": 8.902}, "gpt-6.1-sol/medium": {"attempts": 12, "complete": 12, "first_stdout_p50_s": 0.922, "first_stdout_p90_s": 1.149, "wall_p50_s": 6.973, "wall_p90_s": 7.958}}, "request_attempts_n": 77, "scheduled_capacity_n": 36, "terminal_records_n": 77, "uncensored_completed_n": 76}, "round_id": "llm-latency-short", "round_page": "rounds/llm-latency-short/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep schedule capacity, started requests, terminal requests, uncensored completions, and censored/unstarted requests separate; quantiles use completed requests only.", "arm": null, "claim_id": "CL-WA-llm-latency-long-request-counts-and-endpoints", "claim_status": "DESCRIPTIVE", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public raw logs or measurement definition", "n": 24, "population": "maximum_planned_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 18, "population": "request_attempts_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 18, "population": "terminal_records_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 17, "population": "uncensored_completed_n"}, {"evidence_status": "reproducible from public raw logs or measurement definition", "n": 6, "population": "unstarted_n"}], "evidence": [{"path": "rounds/llm-latency-long/raw/round2-eu-20261006T014516Z.jsonl", "selector": "event rows; numeric client-visible timing only"}, {"path": "rounds/llm-latency-long/reproduce.py", "selector": "summary calculation"}, {"path": "rounds/llm-latency-long/MEASURED.md", "selector": "scheduled and observed populations"}], "evidence_status": "reproducible from public raw logs", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Client-visible timing is not provider transport latency, task success, or a Kogen Build efficacy result.", "metric": "client_visible_latency", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"all_completed_idle_gap_over_90s_n": 17, "collector_fault_n": 1, "completed_600s_or_1200s_crossings_n": 0, "completed_wall_range_s": [112.899, 469.428], "first_stdout_max_s": 1.223, "graded_n": 0, "maximum_planned_n": 24, "request_attempts_n": 18, "terminal_records_n": 18, "uncensored_completed_n": 17, "unstarted_n": 6}, "round_id": "llm-latency-long", "round_page": "rounds/llm-latency-long/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-intent-coverage-audit-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 33, "population": "sampled failed cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 33, "population": "passing controls"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 24, "population": "attributable occurrences"}], "evidence": [{"path": "rounds/intent-coverage-audit/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_intent_coverage_audit", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"additional_failed_test_occurrences_without_output_n": 16, "attributable_failed_test_occurrences_n": 24, "control_cells_n": 33, "control_text_coverage_pct": 88.1, "dropped_or_weakened_n": 8, "failed_cells_n": 33, "failure_text_coverage_pct": 87.9, "hidden_fail_cells_without_output_n": 7, "present_but_failed_n": 15, "qualified_omission_attributions_n": 2, "stricter_than_request_n": 1}, "round_id": "intent-coverage-audit", "round_page": "rounds/intent-coverage-audit/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-overstrict-replay-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 1160, "population": "replayed test cases"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 23, "population": "provided R71 subset"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 140, "population": "cells without source"}], "evidence": [{"path": "rounds/overstrict-replay/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_overstrict_test_replay", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"cells_without_recoverable_test_source_n": 140, "consistent_n": 992, "file_diff_pairs_n": 1638, "indeterminate_n": 0, "logical_test_files_n": 226, "overstrict_n": 168, "overstrict_pct": 14.5, "provided_r71_consistent_n": 14, "provided_r71_overstrict_n": 9, "provided_r71_test_denominator": 23, "test_cases_n": 1160}, "round_id": "overstrict-replay", "round_page": "rounds/overstrict-replay/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-candidate-counterfactual-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 70, "population": "eligible cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 7, "population": "saved candidate files"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 6, "population": "overblocked cells"}], "evidence": [{"path": "rounds/candidate-counterfactual/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "saved_candidate_counterfactual", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"calibration_denominator": 6, "calibration_matches_n": 6, "eligible_finished_scored_cells_n": 70, "genuine_misses_among_overblocked_n": 0, "landed_n": 48, "no_candidate_n": 16, "overblocked_n": 6, "saved_candidate_files_graded_n": 7, "saved_candidate_passes_n": 7, "separate_r67_authorization_adapter_exclusions_n": 60}, "round_id": "candidate-counterfactual", "round_page": "rounds/candidate-counterfactual/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-offline-review-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 613, "population": "confirmed-green events"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 427, "population": "reviewed cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 272, "population": "reproduced revise verdicts"}], "evidence": [{"path": "rounds/offline-review/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "historical_offline_review_proxy", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"confirmed_green_events_n": 613, "distinct_failed_cells_with_unacted_reproduction_n": 39, "false_revise_proxies_n": 40, "related_failure_proxies_n": 57, "reproduced_accept_n": 198, "reproduced_revise_n": 272, "reviewed_cells_n": 427, "revise_denominator": 272}, "round_id": "offline-review", "round_page": "rounds/offline-review/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-offline-review3-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 80, "population": "frozen patches"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 63, "population": "completed paired first runs"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 31, "population": "hidden-pass patches"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 32, "population": "hidden-fail patches"}], "evidence": [{"path": "rounds/offline-review3/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "frozen_offline_reviewer_comparison", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"budget_usd": 10, "executable_cost_usd_mean": 0.0776, "executable_false_veto_den": 31, "executable_false_veto_num": 8, "executable_precision_den": 29, "executable_precision_num": 21, "executable_recall_den": 32, "executable_recall_num": 21, "executable_time_s_mean": 104.7, "final_ledger_usd": 9.5511, "frozen_patches_n": 80, "hidden_fail_patches_n": 32, "hidden_pass_patches_n": 31, "paired_first_runs_complete_n": 63, "toolless_cost_usd_mean": 0.0288, "toolless_false_veto_den": 31, "toolless_false_veto_num": 4, "toolless_precision_den": 23, "toolless_precision_num": 19, "toolless_recall_den": 32, "toolless_recall_num": 19, "toolless_time_s_mean": 31.1}, "round_id": "offline-review3", "round_page": "rounds/offline-review3/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-offline-contract-audit-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 40, "population": "audit calls"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "recoverable contracts"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 5, "population": "known defect cases"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 10, "population": "clean contracts per model"}], "evidence": [{"path": "rounds/offline-contract-audit/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_contract_audit", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"audits_n": 40, "known_false_accept_cases_n": 5, "luna_clean_contract_flags_den": 10, "luna_clean_contract_flags_num": 6, "luna_cost_usd_per_audit_range": [0.0042, 0.0044], "luna_exact_defects_den": 5, "luna_exact_defects_num": 0, "recoverable_contracts_n": 20, "sol_clean_contract_flags_den": 10, "sol_clean_contract_flags_num": 10, "sol_cost_usd_per_audit_range": [0.049, 0.05], "sol_exact_defects_den": 5, "sol_exact_defects_num": 1}, "round_id": "offline-contract-audit", "round_page": "rounds/offline-contract-audit/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-offline-aa-bo3-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 34, "population": "false-promotion simulation tasks"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 17, "population": "selector comparison tasks"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 84, "population": "selector repetitions"}], "evidence": [{"path": "rounds/offline-aa-bo3/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_false_promotion_and_selector_replay", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"deterministic_selector_pass_at_3_pct": 23.5, "direct_luna_pass_at_1_pct": 33.3, "eligible_repetition_denominator": 84, "eligible_repetitions_with_unknown_visible_exit_n": 75, "false_promotion_3_of_3_vs_1_of_3_pct": 3.17, "false_promotion_gap_2_passes_pct": 7.79, "false_promotion_gap_3_passes_pct": 2.04, "false_promotion_gap_4_of_5_vs_1_of_5_pct": 0.95, "historical_tasks_n": 34, "oracle_pass_at_3_pct": 47.1, "reps_per_arm_for_20pp_effect_n": 85, "reps_per_arm_for_30pp_effect_n": 41, "selector_tasks_n": 17}, "round_id": "offline-aa-bo3", "round_page": "rounds/offline-aa-bo3/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-jev-route-ctx-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 18, "population": "route tasks"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 54, "population": "repeated decisions"}], "evidence": [{"path": "rounds/jev-route-ctx/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_route_agreement", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"jev_reference_denominator": 18, "jev_reference_matches": 10, "large_task_misses_n": 3, "repeated_decision_denominator": 54, "repeated_decisions_agree_n": 54, "sol_reference_denominator": 18, "sol_reference_matches": 11, "task_count": 18}, "round_id": "jev-route-ctx", "round_page": "rounds/jev-route-ctx/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-jev-triage-ctx-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 60, "population": "paired sample cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 712, "population": "planned full-pass cells not run"}], "evidence": [{"path": "rounds/jev-triage-ctx/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "No prospective live efficacy result is established by these records.", "metric": "offline_failure_triage", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"agreement_pct": 81.7, "bare_unclear_pct": 41.7, "paired_cells_n": 60, "planned_full_pass_n": 712, "planned_full_pass_ran": false, "prepared_unclear_n": 27, "prepared_unclear_pct": 45.0, "progression_gate_pct": 20}, "round_id": "jev-triage-ctx", "round_page": "rounds/jev-triage-ctx/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-t98-cache-smoke-reported-observation", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 1, "population": "smoke cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 20, "population": "develop requests"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 18, "population": "requests with usage"}], "evidence": [{"path": "rounds/t98-cache-smoke/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H115"], "interpretation": "The public T98 smoke is not denominator- or epoch-matched to historical Kogen/Codex cache-hit summaries and does not test their cost explanation.", "metric": "cache_telemetry_smoke", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"develop_requests_n": 20, "gate_met": false, "gate_threshold_pct": 95, "smoke_cells_n": 1, "turn3_plus_min_cache_hit_pct": 0, "usage_records_n": 18, "weighted_cache_hit_pct": 67.0}, "round_id": "t98-cache-smoke", "round_page": "rounds/t98-cache-smoke/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Offline analysis units, reviewer proxies, candidate counterfactuals, requests, and live Build cells are distinct populations; do not treat these as official Build outcomes.", "arm": null, "claim_id": "CL-WA-t100-cache-lane-reported-observation", "claim_status": "NOT-RUN", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "started cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "finished cells"}, {"evidence_status": "source-reported, not reproducible from public data", "n": 0, "population": "graded cells"}], "evidence": [{"path": "rounds/t100-cache-lane/README.md", "selector": "reported result and reproduction limits"}, {"path": "rounds/README.md", "selector": "round register and status"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H116"], "interpretation": "NOT-RUN; no prospective cache-change measurement is available to resolve the reported percentage change.", "metric": "cache_lane_status", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"finished_n": 0, "graded_n": 0, "itt_n": 0, "planned_n": null, "started_n": 0, "status": "NOT-RUN"}, "round_id": "t100-cache-lane", "round_page": "rounds/t100-cache-lane/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r53-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 110, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 106, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 108, "population": "earlier source-reported deliveries", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r53"}, "path": "results/run-records/r53.jsonl"}, {"filter": {"round": "r53"}, "path": "results/cells.jsonl"}, {"path": "rounds/r53/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 9, "pass": 97, "unknown": 4}, "captured_delivery_rows": 110, "official_outcome_rows": 106, "official_outcomes": {"fail": 10, "pass": 96}}, "round_id": "r53", "round_page": "rounds/r53/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r53b-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 108, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 47, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 35, "population": "older attempts invalidated before controlled relaunch", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r53b"}, "path": "results/run-records/r53b.jsonl"}, {"filter": {"round": "r53b"}, "path": "results/cells.jsonl"}, {"path": "rounds/r53b/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 3, "pass": 44, "unknown": 61}, "captured_delivery_rows": 108, "official_outcome_rows": 47, "official_outcomes": {"fail": 3, "pass": 44}}, "round_id": "r53b", "round_page": "rounds/r53b/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r56-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 222, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 200, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 89, "population": "source-reported P-full comparison cells", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r56"}, "path": "results/run-records/r56.jsonl"}, {"filter": {"round": "r56"}, "path": "results/cells.jsonl"}, {"path": "rounds/r56/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 44, "pass": 156, "unknown": 22}, "captured_delivery_rows": 222, "official_outcome_rows": 200, "official_outcomes": {"fail": 21, "pass": 155, "unresolved": 24}}, "round_id": "r56", "round_page": "rounds/r56/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r56b-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 312, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 226, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r56b"}, "path": "results/run-records/r56b.jsonl"}, {"filter": {"round": "r56b"}, "path": "results/cells.jsonl"}, {"path": "rounds/r56b/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 19, "pass": 207, "unknown": 86}, "captured_delivery_rows": 312, "official_outcome_rows": 226, "official_outcomes": {"fail": 19, "pass": 207}}, "round_id": "r56b", "round_page": "rounds/r56b/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r56c-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 128, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 127, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r56c"}, "path": "results/run-records/r56c.jsonl"}, {"filter": {"round": "r56c"}, "path": "results/cells.jsonl"}, {"path": "rounds/r56c/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 32, "pass": 95, "unknown": 1}, "captured_delivery_rows": 128, "official_outcome_rows": 127, "official_outcomes": {"fail": 32, "pass": 95}}, "round_id": "r56c", "round_page": "rounds/r56c/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r56d-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 37, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 14, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r56d"}, "path": "results/run-records/r56d.jsonl"}, {"filter": {"round": "r56d"}, "path": "results/cells.jsonl"}, {"path": "rounds/r56d/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 2, "pass": 12, "unknown": 23}, "captured_delivery_rows": 37, "official_outcome_rows": 14, "official_outcomes": {"fail": 2, "pass": 12}}, "round_id": "r56d", "round_page": "rounds/r56d/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r56p2-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 336, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 321, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r56p2"}, "path": "results/run-records/r56p2.jsonl"}, {"filter": {"round": "r56p2"}, "path": "results/cells.jsonl"}, {"path": "rounds/r56p2/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 154, "pass": 167, "unknown": 15}, "captured_delivery_rows": 336, "official_outcome_rows": 321, "official_outcomes": {"fail": 122, "pass": 167, "unresolved": 32}}, "round_id": "r56p2", "round_page": "rounds/r56p2/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r57-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 41, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 41, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 312, "population": "source-reported pooled snapshot scored cells", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r57"}, "path": "results/run-records/r57.jsonl"}, {"filter": {"round": "r57"}, "path": "results/cells.jsonl"}, {"path": "rounds/r57/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 7, "pass": 34}, "captured_delivery_rows": 41, "official_outcome_rows": 41, "official_outcomes": {"fail": 7, "not_scored": 1, "pass": 33}}, "round_id": "r57", "round_page": "rounds/r57/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r57b-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 96, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 96, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r57b"}, "path": "results/run-records/r57b.jsonl"}, {"filter": {"round": "r57b"}, "path": "results/cells.jsonl"}, {"path": "rounds/r57b/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 33, "pass": 63}, "captured_delivery_rows": 96, "official_outcome_rows": 96, "official_outcomes": {"fail": 33, "pass": 63}}, "round_id": "r57b", "round_page": "rounds/r57b/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r57c-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 80, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 80, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r57c"}, "path": "results/run-records/r57c.jsonl"}, {"filter": {"round": "r57c"}, "path": "results/cells.jsonl"}, {"path": "rounds/r57c/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 14, "pass": 66}, "captured_delivery_rows": 80, "official_outcome_rows": 80, "official_outcomes": {"fail": 14, "pass": 66}}, "round_id": "r57c", "round_page": "rounds/r57c/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r57d-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 97, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 96, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r57d"}, "path": "results/run-records/r57d.jsonl"}, {"filter": {"round": "r57d"}, "path": "results/cells.jsonl"}, {"path": "rounds/r57d/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 33, "pass": 63, "unknown": 1}, "captured_delivery_rows": 97, "official_outcome_rows": 96, "official_outcomes": {"fail": 33, "pass": 63}}, "round_id": "r57d", "round_page": "rounds/r57d/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r57e-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 80, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 78, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r57e"}, "path": "results/run-records/r57e.jsonl"}, {"filter": {"round": "r57e"}, "path": "results/cells.jsonl"}, {"path": "rounds/r57e/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 23, "pass": 55, "unknown": 2}, "captured_delivery_rows": 80, "official_outcome_rows": 78, "official_outcomes": {"fail": 23, "pass": 55}}, "round_id": "r57e", "round_page": "rounds/r57e/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r64-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 51, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 38, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 85, "population": "source-reported historical plan-shell denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 85, "population": "source-reported historical Sol-high denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 68, "population": "source-reported historical Luna-max denominator", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r64"}, "path": "results/run-records/r64.jsonl"}, {"filter": {"round": "r64"}, "path": "results/cells.jsonl"}, {"path": "rounds/r64/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 6, "pass": 32, "unknown": 13}, "captured_delivery_rows": 51, "official_outcome_rows": 38, "official_outcomes": {"fail": 6, "pass": 32}}, "round_id": "r64", "round_page": "rounds/r64/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r64b-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 68, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 61, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r64b"}, "path": "results/run-records/r64b.jsonl"}, {"filter": {"round": "r64b"}, "path": "results/cells.jsonl"}, {"path": "rounds/r64b/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 19, "pass": 40, "unknown": 9}, "captured_delivery_rows": 68, "official_outcome_rows": 61, "official_outcomes": {"fail": 19, "pass": 40, "unresolved": 2}}, "round_id": "r64b", "round_page": "rounds/r64b/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r64c-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 102, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 90, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r64c"}, "path": "results/run-records/r64c.jsonl"}, {"filter": {"round": "r64c"}, "path": "results/cells.jsonl"}, {"path": "rounds/r64c/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 31, "pass": 55, "unknown": 16}, "captured_delivery_rows": 102, "official_outcome_rows": 90, "official_outcomes": {"fail": 31, "pass": 55, "unresolved": 4}}, "round_id": "r64c", "round_page": "rounds/r64c/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r64d-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 20, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 19, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r64d"}, "path": "results/run-records/r64d.jsonl"}, {"filter": {"round": "r64d"}, "path": "results/cells.jsonl"}, {"path": "rounds/r64d/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 5, "pass": 14, "unknown": 1}, "captured_delivery_rows": 20, "official_outcome_rows": 19, "official_outcomes": {"fail": 5, "pass": 14}}, "round_id": "r64d", "round_page": "rounds/r64d/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r64e-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 34, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 27, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}], "evidence": [{"filter": {"round_id": "r64e"}, "path": "results/run-records/r64e.jsonl"}, {"filter": {"round": "r64e"}, "path": "results/cells.jsonl"}, {"path": "rounds/r64e/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 9, "pass": 17, "unknown": 8}, "captured_delivery_rows": 34, "official_outcome_rows": 27, "official_outcomes": {"fail": 9, "pass": 17, "unresolved": 1}}, "round_id": "r64e", "round_page": "rounds/r64e/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r69-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 459, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 95, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 400, "population": "predeclared design cells", "source_status": "source-reported, not reproducible from public data"}, {"n": 396, "population": "source-reported resolved ITT total across four arms", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r69"}, "path": "results/run-records/r69.jsonl"}, {"filter": {"round": "r69"}, "path": "results/cells.jsonl"}, {"path": "rounds/r69/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"fail": 37, "pass": 113, "unknown": 309}, "captured_delivery_rows": 459, "official_outcome_rows": 150, "official_outcomes": {"fail": 37, "pass": 113}}, "round_id": "r69", "round_page": "rounds/r69/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r70-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 6, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 0, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 350, "population": "phase-A admitted design slots", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r70"}, "path": "results/run-records/r70.jsonl"}, {"filter": {"round": "r70"}, "path": "results/cells.jsonl"}, {"path": "rounds/r70/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {"unknown": 6}, "captured_delivery_rows": 6, "official_outcome_rows": 0, "official_outcomes": {}}, "round_id": "r70", "round_page": "rounds/r70/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record and official cell rows separately by exact round ID; report invalid control_apply applications separately from model outcomes. Only exact shared cell IDs support row-level linkage.", "arm": null, "claim_id": "CL-WA-r71-public-export-coverage", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 50, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 18, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 48, "population": "source-reported planned scored cells", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r71"}, "path": "results/run-records/r71.jsonl"}, {"filter": {"round": "r71"}, "path": "results/cells.jsonl"}, {"path": "rounds/r71/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "All 49 graded captured IDs have exact official-export rows. The export contains 40 model outcomes and 9 invalid control applications; the model outcome rows do not complete the planned 48-cell same-task contrast. Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_control_applications": {"invalid": 9}, "captured_delivery_outcomes": {"fail": 3, "pass": 37, "unknown": 1}, "captured_delivery_rows": 50, "official_outcome_rows": 49, "official_outcomes": {"fail": 3, "not_scored": 2, "pass": 35, "unresolved": 9}, "official_result_types": {"fail": 3, "invalid_control": 9, "pass": 37}}, "round_id": "r71", "round_page": "rounds/r71/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count run-record rows and cells rows independently by exact round label. They are separate populations, not a single ITT denominator; do not join without an exact shared cell ID.", "arm": null, "claim_id": "CL-WA-r74-public-export-coverage", "claim_status": "WITHDRAWN", "comparison": null, "denominator": null, "denominator_variants": [{"n": 0, "population": "captured delivery rows in run-record export", "source_status": "reproducible from public data"}, {"n": 0, "population": "official outcome rows in cells export", "source_status": "reproducible from public data"}, {"n": 480, "population": "initial confirmatory specifications", "source_status": "source-reported, not reproducible from public data"}, {"n": 82, "population": "unstarted specifications removed by amendment", "source_status": "source-reported, not reproducible from public data"}, {"n": 398, "population": "retained amended specifications", "source_status": "source-reported, not reproducible from public data"}, {"n": 167, "population": "status snapshot pulled/graded scored records before cancellation", "source_status": "source-reported, not reproducible from public data"}, {"n": 176, "population": "source-captured run-record rows including diagnostics", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"filter": {"round_id": "r74"}, "path": "results/run-records/index.json"}, {"filter": {"round": "r74"}, "path": "results/cells.jsonl"}, {"path": "rounds/r74/README.md", "selector": "public outcome reconciliation status"}], "evidence_status": "reproducible from public data for public export counts; source-reported, not reproducible from public data for historical/planned denominator variants", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Row-count agreement or disagreement is not evidence of a performance effect.", "metric": "separate_public_export_row_counts", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"captured_delivery_outcomes": {}, "captured_delivery_rows": 0, "official_outcome_rows": 0, "official_outcomes": {}}, "round_id": "r74", "round_page": "rounds/r74/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Use the exact cohort mapping in VERIFICATION.md, count unresolved outcomes in the row denominator, and divide available recorded usd_est by export passes; do not substitute historical rates or pool other rounds.", "arm": "Plan-shell", "claim_id": "CL-WA-rails-catalogue-plan-shell-public-export-vs-reported", "claim_status": "INTERIM", "comparison": null, "denominator": 65, "denominator_variants": [{"n": 65, "population": "selected public export rows", "source_status": "reproducible from public data"}, {"n": 85, "population": "historical source-reported arm denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 0.17637, "population": "source-reported historical cost per pass", "source_status": "source-reported, not reproducible from public data", "unit": "USD/pass"}], "evidence": [{"filter": {"arm_map": "rounds/rails-catalogue/VERIFICATION.md COHORTS", "rounds": ["r60", "r60b", "r63", "r63b", "r64", "r64b", "r64c"], "task_ids": "rounds/rails-catalogue/README.md task list"}, "path": "results/cells.jsonl"}, {"path": "rounds/rails-catalogue/VERIFICATION.md", "selector": "reproduction code and output"}], "evidence_status": "reproducible from public data for export counts/cost arithmetic; source-reported, not reproducible from public data for historical denominators and cost claims", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The retrospective aggregate remains INTERIM and does not establish held-out superiority.", "metric": "retrospective_rails_export_pass_rate_and_denominator", "numerator": 52, "publication_status": "STATUS_ONLY", "reported_values": {"export_cost_per_pass_usd": 0.119077, "historical_source_reported_cost_per_pass_usd": 0.17637, "historical_source_reported_denominator": 85, "public_export_passes": 52, "public_export_rows": 65, "recorded_cost_sum_usd": 6.192018}, "round_id": "rails-catalogue", "round_page": "rounds/rails-catalogue/README.md", "schema_version": "1.0", "unit": "selected public cells"}
{"analysis_rule": "Use the exact cohort mapping in VERIFICATION.md, count unresolved outcomes in the row denominator, and divide available recorded usd_est by export passes; do not substitute historical rates or pool other rounds.", "arm": "Sol high", "claim_id": "CL-WA-rails-catalogue-sol-high-public-export-vs-reported", "claim_status": "INTERIM", "comparison": null, "denominator": 81, "denominator_variants": [{"n": 81, "population": "selected public export rows", "source_status": "reproducible from public data"}, {"n": 85, "population": "historical source-reported arm denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 0.793775, "population": "source-reported historical cost per pass", "source_status": "source-reported, not reproducible from public data", "unit": "USD/pass"}], "evidence": [{"filter": {"arm_map": "rounds/rails-catalogue/VERIFICATION.md COHORTS", "rounds": ["r60", "r60b", "r63", "r63b", "r64", "r64b", "r64c"], "task_ids": "rounds/rails-catalogue/README.md task list"}, "path": "results/cells.jsonl"}, {"path": "rounds/rails-catalogue/VERIFICATION.md", "selector": "reproduction code and output"}], "evidence_status": "reproducible from public data for export counts/cost arithmetic; source-reported, not reproducible from public data for historical denominators and cost claims", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The retrospective aggregate remains INTERIM and does not establish held-out superiority.", "metric": "retrospective_rails_export_pass_rate_and_denominator", "numerator": 50, "publication_status": "STATUS_ONLY", "reported_values": {"export_cost_per_pass_usd": 0.680384, "historical_source_reported_cost_per_pass_usd": 0.793775, "historical_source_reported_denominator": 85, "public_export_passes": 50, "public_export_rows": 81, "recorded_cost_sum_usd": 34.019188}, "round_id": "rails-catalogue", "round_page": "rounds/rails-catalogue/README.md", "schema_version": "1.0", "unit": "selected public cells"}
{"analysis_rule": "Use the exact cohort mapping in VERIFICATION.md, count unresolved outcomes in the row denominator, and divide available recorded usd_est by export passes; do not substitute historical rates or pool other rounds.", "arm": "Luna max", "claim_id": "CL-WA-rails-catalogue-luna-max-public-export-vs-reported", "claim_status": "INTERIM", "comparison": null, "denominator": 64, "denominator_variants": [{"n": 64, "population": "selected public export rows", "source_status": "reproducible from public data"}, {"n": 68, "population": "historical source-reported arm denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 0.122729, "population": "source-reported historical cost per pass", "source_status": "source-reported, not reproducible from public data", "unit": "USD/pass"}], "evidence": [{"filter": {"arm_map": "rounds/rails-catalogue/VERIFICATION.md COHORTS", "rounds": ["r60", "r60b", "r63", "r63b", "r64", "r64b", "r64c"], "task_ids": "rounds/rails-catalogue/README.md task list"}, "path": "results/cells.jsonl"}, {"path": "rounds/rails-catalogue/VERIFICATION.md", "selector": "reproduction code and output"}], "evidence_status": "reproducible from public data for export counts/cost arithmetic; source-reported, not reproducible from public data for historical denominators and cost claims", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The retrospective aggregate remains INTERIM and does not establish held-out superiority.", "metric": "retrospective_rails_export_pass_rate_and_denominator", "numerator": 27, "publication_status": "STATUS_ONLY", "reported_values": {"export_cost_per_pass_usd": 0.071076, "historical_source_reported_cost_per_pass_usd": 0.122729, "historical_source_reported_denominator": 68, "public_export_passes": 27, "public_export_rows": 64, "recorded_cost_sum_usd": 1.919042}, "round_id": "rails-catalogue", "round_page": "rounds/rails-catalogue/README.md", "schema_version": "1.0", "unit": "selected public cells"}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r53-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 28, "population": "staged_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 36, "population": "staged_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 32, "population": "direct_luna_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 36, "population": "direct_luna_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 34, "population": "direct_sol_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 36, "population": "direct_sol_denominator", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"path": "rounds/r53/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Historical R53 result summary; the public page withholds the comparison because the official and captured exports conflict.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"direct_luna_denominator": 36, "direct_luna_passes": 32, "direct_sol_denominator": 36, "direct_sol_passes": 34, "staged_denominator": 36, "staged_passes": 28}, "round_id": "r53", "round_page": "rounds/r53/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r56-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 89, "population": "p_full_source_reported_n", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"path": "rounds/r56/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Historical stage-ablation summary; public rows do not reproduce its filtered analysis.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"p_full_source_reported_n": 89, "remove_context_delta_pp": 0.3, "remove_contract_delta_pp": 6.53, "remove_plan_delta_pp": 11.55, "remove_review_delta_pp": 1.04}, "round_id": "r56", "round_page": "rounds/r56/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r57-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported planned population; one slot has no recovered cell ID or outcome", "n": 312, "population": "historical_pooled_cells", "source_status": "source-reported planned population; one slot has no recovered cell ID or outcome"}, {"evidence_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered 311 rows", "n": 311, "population": "historical_pooled_graded", "source_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered 311 rows"}, {"evidence_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort", "n": 114, "population": "shell_passes", "source_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort"}, {"evidence_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort", "n": 155, "population": "shell_denominator", "source_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort"}, {"evidence_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort", "n": 111, "population": "default_passes", "source_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort"}, {"evidence_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort", "n": 156, "population": "default_denominator", "source_status": "reproducible from public L0 reconciliation CSV and public outcome export for the recovered cohort"}], "evidence": [{"path": "rounds/r57/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}, {"path": "rounds/l0-reconcile/data/r57-pooled-itt.csv", "selector": "311 recovered pooled ITT rows"}, {"path": "reproduce/reconcile_l0.py", "selector": "r57()"}], "evidence_status": "Reproducible from the public L0 reconciliation CSV and public outcome export for the recovered 311-cell pooled population and published interim analysis. One planned Studio shell-only slot has no recovered cell ID or outcome, so the full planned population remains scoped and incomplete. INTERIM. The complete planned 312-cell population remains source-reported, not reproducible from public data.", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The L0 CSV reproduces the 311 recovered pooled outcomes and published interim bound. One planned Studio shell-only slot remains missing; these reconstructed results do not establish a general tools effect or a complete 312-cell ITT cohort.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reconciliation_status": "Public L0 CSV analysis is reproducible; official outcome join status remains explicitly scoped in evidence_status.", "reported_values": {"default_denominator": 156, "default_passes": 111, "historical_pooled_cells": 312, "historical_pooled_graded": 311, "shell_denominator": 155, "shell_passes": 114}, "round_id": "r57", "round_page": "rounds/r57/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r64-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 52, "population": "plan_shell_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 85, "population": "plan_shell_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 50, "population": "sol_high_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 85, "population": "sol_high_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 27, "population": "luna_max_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 68, "population": "luna_max_denominator", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"path": "rounds/r64/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "Historical Rails catalogue numbers are not reproduced by public export.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"luna_max_denominator": 68, "luna_max_passes": 27, "plan_shell_denominator": 85, "plan_shell_passes": 52, "plan_shell_vs_luna_p": 0.000329, "plan_shell_vs_sol_p": 0.403743, "sol_high_denominator": 85, "sol_high_passes": 50}, "round_id": "r64", "round_page": "rounds/r64/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r69-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 63, "population": "spec_passes", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 99, "population": "spec_denominator", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 66, "population": "approach_passes", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 99, "population": "approach_denominator", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 66, "population": "acceptance_passes", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 98, "population": "acceptance_denominator", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 59, "population": "none_passes", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 100, "population": "none_denominator", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 396, "population": "resolved_itt_n", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}, {"evidence_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial", "n": 400, "population": "planned_n", "source_status": "reproducible from public L0 reconstruction CSV; public outcome join is partial"}], "evidence": [{"path": "rounds/r69/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}, {"path": "rounds/l0-reconcile/data/r69-itt.csv", "selector": "400 assigned rows with inclusion and public-grade-match fields"}, {"path": "reproduce/reconcile_l0.py", "selector": "r69()"}], "evidence_status": "Reproducible from the public L0 reconciliation CSV for the recovered 400-row assignment, 396 included ITT rows, arm counts, and intervals. The reconstruction has 327 internal official grades, of which 150 now match the refreshed public official outcome export; the official join remains partial. INTERIM. The full historical ITT partition remains source-reported, not reproducible from public data.", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "L0 reproduces the selected R69 cohort from its published CSV, but only 150/327 internal grades match the canonical public official outcome export. Retain INTERIM: the registered rule requires 400 official grades and a gain of at least 15 percentage points with one-sided task-stratified p < 0.05; the refreshed public cohort does not establish that comparison.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reconciliation_status": "Public L0 CSV analysis is reproducible; the refreshed official export verifies 150/327 internal grades, and the registered 400-grade decision analysis remains incomplete.", "reported_values": {"acceptance_denominator": 98, "acceptance_passes": 66, "approach_denominator": 99, "approach_passes": 66, "none_denominator": 100, "none_passes": 59, "planned_n": 400, "resolved_itt_n": 396, "spec_denominator": 99, "spec_passes": 63}, "round_id": "r69", "round_page": "rounds/r69/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r70-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 40, "population": "original_rve_scored_n", "source_status": "source-reported, not reproducible from public data"}, {"n": 5, "population": "extra_ungraded_deliveries_n", "source_status": "source-reported, not reproducible from public data"}, {"n": 15, "population": "rerun_n", "source_status": "source-reported, not reproducible from public data"}, {"n": 29, "population": "extension_n", "source_status": "source-reported, not reproducible from public data"}, {"n": 4, "population": "task8_smoke_n", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"path": "rounds/r70/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "R70 subcohorts remain separate; the 350-slot phase-A design is not closed by the RvE receipts.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"extension_n": 29, "extra_ungraded_deliveries_n": 5, "original_rve_scored_n": 40, "rerun_n": 15, "task8_smoke_n": 4}, "round_id": "r70", "round_page": "rounds/r70/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r71-source-reported-denominator-variants", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"n": 21, "population": "provided_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 24, "population": "provided_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 15, "population": "shaped_passes", "source_status": "source-reported, not reproducible from public data"}, {"n": 24, "population": "shaped_denominator", "source_status": "source-reported, not reproducible from public data"}, {"n": 48, "population": "planned_n", "source_status": "source-reported, not reproducible from public data"}, {"n": 50, "population": "captured_deliveries_n", "source_status": "source-reported, not reproducible from public data"}], "evidence": [{"path": "rounds/r71/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The refreshed export verifies all 49 graded captured IDs, comprising 40 model outcomes and 9 invalid controls. Eight of the 48 planned model slots remain without model results, so the historical provided-versus-shaped paired comparison remains unreconciled.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"captured_deliveries_n": 50, "planned_n": 48, "provided_denominator": 24, "provided_passes": 21, "shaped_denominator": 24, "shaped_passes": 15}, "round_id": "r71", "round_page": "rounds/r71/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical source denominators separate from public export counts and from adjacent amendments or replications.", "arm": null, "claim_id": "CL-WA-r74-source-reported-denominator-variants", "claim_status": "WITHDRAWN", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "reproducible from public L0 coverage CSV; source provenance remains reported", "n": 480, "population": "initial_n", "source_status": "reproducible from public L0 coverage CSV; source provenance remains reported"}, {"evidence_status": "reproducible from public L0 coverage CSV; source provenance remains reported", "n": 82, "population": "removed_unstarted_n", "source_status": "reproducible from public L0 coverage CSV; source provenance remains reported"}, {"evidence_status": "reproducible from public L0 coverage CSV; source provenance remains reported", "n": 398, "population": "retained_n", "source_status": "reproducible from public L0 coverage CSV; source provenance remains reported"}, {"evidence_status": "reproducible from public L0 CSV; not independently joined to canonical official outcome export", "n": 167, "population": "snapshot_scored_graded_n", "source_status": "reproducible from public L0 CSV; not independently joined to canonical official outcome export"}, {"evidence_status": "reproducible from public L0 coverage CSV; source provenance remains reported", "n": 176, "population": "snapshot_run_record_n", "source_status": "reproducible from public L0 coverage CSV; source provenance remains reported"}, {"evidence_status": "reproducible from public L0 coverage CSV; no tagged public outcome rows", "n": 0, "population": "public_capture_n", "source_status": "reproducible from public L0 coverage CSV; no tagged public outcome rows"}, {"evidence_status": "reproducible from public L0 coverage CSV; no tagged public outcome rows", "n": 0, "population": "public_official_rows_n", "source_status": "reproducible from public L0 coverage CSV; no tagged public outcome rows"}], "evidence": [{"path": "rounds/r74/README.md", "selector": "status and source reconciliation note"}, {"reproducibility": "source-reported, not reproducible from public data", "source": "research synthesis summary"}, {"path": "rounds/l0-reconcile/data/r74-graded-subset.csv", "selector": "167 recovered internal scored-grade rows"}, {"path": "rounds/l0-reconcile/data/r74-smoke-attempts.csv", "selector": "9 smoke attempt grade rows"}, {"path": "rounds/l0-reconcile/data/r74-coverage.csv", "selector": "retained, missing, and public-export counts"}, {"path": "reproduce/reconcile_l0.py", "selector": "r74()"}], "evidence_status": "Reproducible from public L0 reconciliation CSVs for the recovered 167 internal scored rows, 9 smoke attempts, and coverage counts. These rows are not independently joined to the canonical official outcome export; 231 retained slots lack recovered cell IDs and terminal statuses. WITHDRAWN. 231 retained slots remain source-reported, not reproducible from public data.", "exact_cell_ids": [], "hypothesis_ids": [], "interpretation": "The L0 CSVs reproduce the selected internal snapshot counts, but the retained cohort remains incomplete and has no public scored outcome join. Preserve WITHDRAWN; do not use the recovered subset as an efficacy estimate.", "metric": "historical_denominator_conflict", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reconciliation_status": "Public L0 CSV analysis is reproducible; official outcome join status remains explicitly scoped in evidence_status.", "reported_values": {"initial_n": 480, "public_capture_n": 0, "public_official_rows_n": 0, "removed_unstarted_n": 82, "retained_n": 398, "snapshot_run_record_n": 176, "snapshot_scored_graded_n": 167}, "round_id": "r74", "round_page": "rounds/r74/README.md", "schema_version": "1.0", "unit": null}
{"analysis_rule": "Retain historical values only as source-reported, not reproducible from public data. The task set, denominator, official grades, and arm mapping are absent; do not infer an advantage.", "arm": null, "claim_id": "CL-H111-historical-staged-and-direct-comparison", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"reproducibility": "source-reported, not reproducible from public data", "source": "historical summary claim; exact public cohort and row-level inputs not identified"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H111"], "interpretation": "The historical summary reports direct-model pass-rate values of 85% and 89%, an approximately 2.5-fold lower cost per pass, and a 1.5-to-2-fold wall-time difference. The comparison direction and cohort cannot be reconstructed.", "metric": "historical_staged_and_direct_comparison_claim", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"cohort_reconstructed": false, "reported_cost_per_pass_factor": 2.5, "reported_direct_pass_rates_pct": [85, 89], "reported_wall_time_factor_range": [1.5, 2]}, "round_id": null, "round_page": null, "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep the summarized task count and each arm result source-reported; the shared-task membership, grade rows, and compatible cost definitions are unavailable.", "arm": null, "claim_id": "CL-H112-historical-route-cost-frontier", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [{"evidence_status": "source-reported, not reproducible from public data", "n": 18, "population": "historically summarized shared tasks"}], "evidence": [{"reproducibility": "source-reported, not reproducible from public data", "source": "historical summary claim; exact public cohort and row-level inputs not identified"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H112"], "interpretation": "Historical values report 76% and $0.38 per pass for direct Sol, 70% and $0.38 for route-v3, 50% and $0.047 for direct Luna max, and four unsolved rails-ft tasks. These values do not establish a current cost-quality frontier.", "metric": "historical_route_cost_frontier_claim", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"direct_luna_max_cost_per_pass_usd": 0.047, "direct_luna_max_pass_rate_pct": 50, "direct_sol_cost_per_pass_usd": 0.38, "direct_sol_pass_rate_pct": 76, "route_v3_cost_per_pass_usd": 0.38, "route_v3_pass_rate_pct": 70, "shared_tasks_n": 18, "unsolved_rails_ft_tasks_n": 4}, "round_id": null, "round_page": null, "schema_version": "1.0", "unit": null}
{"analysis_rule": "Keep historical cache summaries separate from the public T98 smoke; their request counts, epochs, and token definitions are not compatible.", "arm": null, "claim_id": "CL-H115-historical-cache-hit-comparison", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"reproducibility": "source-reported, not reproducible from public data", "source": "historical summary claim; raw request ledgers and compatible token definitions are unavailable"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H115"], "interpretation": "Historical summaries report 26.5% Kogen cache hit, 93% Codex cache hit, and a separate local aggregate near 70%. The epochs and denominators differ or are unavailable; the claimed token-spend explanation is unverified.", "metric": "historical_cache_hit_comparison_claim", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"codex_cache_hit_pct": 93, "kogen_cache_hit_pct": 26.5, "last_value_qualifier": "approximately", "separate_local_aggregate_cache_hit_pct": 70, "token_spend_explanation_verified": false}, "round_id": null, "round_page": null, "schema_version": "1.0", "unit": null}
{"analysis_rule": "Preserve the numeric change as source-reported. The metric definition, baseline, period, and request denominator are unavailable.", "arm": null, "claim_id": "CL-H116-historical-cache-change", "claim_status": "INTERIM", "comparison": null, "denominator": null, "denominator_variants": [], "evidence": [{"reproducibility": "source-reported, not reproducible from public data", "source": "historical summary claim; metric definition and underlying request records not identified"}], "evidence_status": "source-reported, not reproducible from public data", "exact_cell_ids": [], "hypothesis_ids": ["H116"], "interpretation": "The historical summary reports a 19% cache increase after a change, but does not define whether this is a relative or percentage-point change or identify a baseline and sample.", "metric": "historical_cache_change_claim", "numerator": null, "publication_status": "RECONCILIATION_QUEUE", "reported_values": {"baseline": null, "change_basis": null, "period": null, "reported_change_pct": 19, "request_denominator": null}, "round_id": null, "round_page": null, "schema_version": "1.0", "unit": null}
{"analysis_rule": "Count repair_outcome=pass among the six matched repair rows. Treat the registered per-test pass-set preservation condition as unverified; aggregate test counts are only a separate diagnostic.", "claim_id": "CL-L3-REPAIR-RESCUE-RATE", "claim_status": "VALID", "denominator": 6, "evidence": [{"path": "rounds/l3-repair/data/cells.csv", "selector": "six kind=pair rows; repair_cell_id and repair_outcome"}, {"path": "rounds/l3-repair/DECISION-RULE.md", "selector": "rescue threshold and per-test preservation rule"}], "evidence_status": "Reproducible from six published aggregate repair rows; individual test-set preservation is unverifiable.", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-2-elixir-l3r31__r1", "codex__gpt-6-luna__max__default__r70-2-elixir-l3r33__r1", "codex__gpt-6-luna__max__default__r70-2-go-l3r32__r1", "codex__gpt-6-luna__max__default__r70-5-go-l3r1__r1", "codex__gpt-6-luna__max__default__r70-5-go-l3r32__r1", "codex__gpt-6-luna__max__default__r70-7-rust-l3r31__r1"], "hypothesis_ids": ["H12"], "interpretation": "Two of six repairs rescued the official hidden-suite outcome; the three-rescue threshold was not met. The result does not show a causal repair benefit or verify per-test no-regression.", "metric": "repair_rescue_count", "numerator": 2, "publication_status": "STATUS_ONLY", "reported_values": {"per_test_pass_set_preservation": "unverifiable", "repairs": 6, "rescued": 2, "threshold_met": false, "threshold_rescues": 3}, "round_id": "l3-repair", "round_page": "rounds/l3-repair/README.md", "schema_version": "1.0", "unit": "matched failed implementation repairs"}
{"analysis_rule": "Score valid auditor verdicts against the original source-case outcomes. Report precision and false-demotion rate by auditor model against the preregistered targets; count timeout separately from a valid verdict.", "claim_id": "CL-L6-AUDITOR-ADVISORY-REPLAY", "claim_status": "VALID", "denominator": null, "evidence": [{"path": "rounds/l6-auditor-replay/data/case_list.csv", "selector": "40 source patches and original outcomes"}, {"path": "rounds/l6-auditor-replay/data/calls.csv", "selector": "80 auditor requests; one timeout"}, {"path": "rounds/l6-auditor-replay/reproduce/score_l6.py", "selector": "public scoring transformation"}], "evidence_status": "Reproducible from the published offline case and call ledgers; not a live Build-gate estimate.", "exact_cell_ids": ["codex__gpt-6-luna__max__default__r70-7-rust__r31", "codex__gpt-6-luna__max__default__r70-2-elixir__r31", "codex__gpt-6.1-sol__medium__default__r70-5-elixir__r3", "codex__gpt-6-luna__max__default__r70-1-go__r16", "codex__gpt-6-luna__max__default__r70-1-ts-bun__r16", "codex__gpt-6-luna__max__default__r70-7-ts-bun__r3", "codex__gpt-6.1-sol__medium__default__r70-6-elixir__r3", "codex__gpt-6-luna__max__default__r70-8-rust__r90", "codex__gpt-6.1-sol__medium__default__r70-6-gleam__r4", "codex__gpt-6-luna__max__default__r70-2-ts-bun__r31", "codex__gpt-6.1-sol__medium__default__r70-6-elixir__r1", "codex__gpt-6-luna__max__default__r70-4-ts-bun-fe2__r9", "codex__gpt-6-luna__max__default__r70-6-elixir__r1", "codex__gpt-6-luna__max__default__r70-4-elixir-fe2__r9", "codex__gpt-6-luna__max__default__r70-5-ts-bun__r31", "codex__gpt-6-luna__max__default__r70-6-go__r1", "codex__gpt-6-luna__max__default__r70-4-rust__r13", "codex__gpt-6-luna__max__default__r70-2-elixir__r32", "codex__gpt-6-luna__max__default__r70-3-rust__r14", "codex__gpt-6-luna__max__default__r70-6-go__r2", "codex__gpt-6-luna__max__default__r70-7-gleam__r1", "codex__gpt-6-luna__max__default__r70-4-go__r20", "codex__gpt-6-luna__max__default__r70-5-go__r1", "codex__gpt-6-luna__max__default__r70-5-gleam__r3", "codex__gpt-6-luna__max__default__r70-5-go__r32", "codex__gpt-6-luna__max__default__r70-4-ts-bun__r23", "codex__gpt-6-luna__max__default__r70-4-go__r19", "codex__gpt-6-luna__max__default__r70-6-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-5-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-5-gleam__r1", "codex__gpt-6-luna__max__default__r70-1-elixir__r6", "codex__gpt-6-luna__max__default__r70-7-gleam__r3", "codex__gpt-6.1-sol__medium__default__r70-6-rust__r2", "codex__gpt-6-luna__max__default__r70-8-go-v2__r1", "codex__gpt-6.1-sol__medium__default__r70-5-go__r1", "codex__gpt-6-luna__max__default__r70-3-elixir__r7", "codex__gpt-6-luna__max__default__r70-5-ts-bun__r1", "codex__gpt-6-luna__max__default__r70-4-ts-bun-fe2__r10", "codex__gpt-6-luna__max__default__r70-7-elixir__r31", "codex__gpt-6-luna__max__default__r70-8-rust-v2__r1"], "hypothesis_ids": ["H67"], "interpretation": "Neither auditor meets the precision and false-demotion targets. The findings support advisory-only use for this sample, not a live gate.", "metric": "offline_auditor_target_check", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"interpretation": "advisory only for this selected offline sample", "luna_max": {"false_demotions_percent": 35.0, "max_false_demotions_percent": 5.0, "precision_target_percent": 90.0, "veto_precision_percent": 56.25}, "requests": 80, "sol_low": {"false_demotions_percent": 25.0, "max_false_demotions_percent": 5.0, "precision_target_percent": 90.0, "veto_precision_percent": 66.67}, "timeout_requests": 1, "valid_verdicts": 79}, "round_id": "l6-auditor-replay", "round_page": "rounds/l6-auditor-replay/README.md", "schema_version": "1.0", "unit": "auditor requests on 40 selected source patches"}
{"analysis_rule": "Within each stack, compute the pass fraction for each of six task identities, then take their unweighted arithmetic mean. Show the raw combined pass totals separately; do not use this post-hoc sensitivity to execute the under-specified four-pass decision threshold.", "claim_id": "CL-SOL-LANG-REPLICATION-POSTHOC-EQUAL-TASK", "claim_status": "VALID", "denominator": null, "evidence": [{"path": "rounds/lang-sol-replication/data/cells.csv", "selector": "all 58 rows; exact cell IDs; grouped by stack and six task identities"}, {"path": "rounds/lang-sol-replication/DECISION-RULE.md", "selector": "registered combined branch and disclosed post-hoc formula"}], "evidence_status": "Reproducible from the published round-local CSV; equal-task formula is post-hoc because the original combined formula was not recorded.", "exact_cell_ids": ["codex__gpt-6.1-sol__medium__default__r70-1-rust__r1", "codex__gpt-6.1-sol__medium__default__r70-1-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-1-rust__r3", "codex__gpt-6.1-sol__medium__default__r70-2-rust__r1", "codex__gpt-6.1-sol__medium__default__r70-2-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-2-rust__r3", "codex__gpt-6.1-sol__medium__default__r70-4-rust-fe2__r1", "codex__gpt-6.1-sol__medium__default__r70-4-rust-fe2__r2", "codex__gpt-6.1-sol__medium__default__r70-4-rust-fe2__r3", "codex__gpt-6.1-sol__medium__default__r70-1-go__r1", "codex__gpt-6.1-sol__medium__default__r70-1-go__r2", "codex__gpt-6.1-sol__medium__default__r70-1-go__r3", "codex__gpt-6.1-sol__medium__default__r70-2-go__r1", "codex__gpt-6.1-sol__medium__default__r70-2-go__r2", "codex__gpt-6.1-sol__medium__default__r70-2-go__r3", "codex__gpt-6.1-sol__medium__default__r70-4-go-fe2__r1", "codex__gpt-6.1-sol__medium__default__r70-4-go-fe2__r2", "codex__gpt-6.1-sol__medium__default__r70-4-go-fe2__r3", "codex__gpt-6.1-sol__medium__default__r70-1-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-1-ts-bun__r2", "codex__gpt-6.1-sol__medium__default__r70-1-ts-bun__r3", "codex__gpt-6.1-sol__medium__default__r70-2-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-2-ts-bun__r2", "codex__gpt-6.1-sol__medium__default__r70-2-ts-bun__r3", "codex__gpt-6.1-sol__medium__default__r70-4-ts-bun-fe2__r1", "codex__gpt-6.1-sol__medium__default__r70-4-ts-bun-fe2__r2", "codex__gpt-6.1-sol__medium__default__r70-4-ts-bun-fe2__r3", "codex__gpt-6.1-sol__medium__default__r70-5-rust__r1", "codex__gpt-6.1-sol__medium__default__r70-5-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-5-rust__r3", "codex__gpt-6.1-sol__medium__default__r70-5-rust__r4", "codex__gpt-6.1-sol__medium__default__r70-6-rust__r1", "codex__gpt-6.1-sol__medium__default__r70-6-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-6-rust__r3", "codex__gpt-6.1-sol__medium__default__r70-6-rust__r4", "codex__gpt-6.1-sol__medium__default__r70-7-rust__r1", "codex__gpt-6.1-sol__medium__default__r70-7-rust__r2", "codex__gpt-6.1-sol__medium__default__r70-7-rust__r3", "codex__gpt-6.1-sol__medium__default__r70-5-go__r1", "codex__gpt-6.1-sol__medium__default__r70-5-go__r2", "codex__gpt-6.1-sol__medium__default__r70-5-go__r3", "codex__gpt-6.1-sol__medium__default__r70-5-go__r4", "codex__gpt-6.1-sol__medium__default__r70-6-go__r1", "codex__gpt-6.1-sol__medium__default__r70-6-go__r2", "codex__gpt-6.1-sol__medium__default__r70-6-go__r3", "codex__gpt-6.1-sol__medium__default__r70-6-go__r4", "codex__gpt-6.1-sol__medium__default__r70-7-go__r1", "codex__gpt-6.1-sol__medium__default__r70-7-go__r2", "codex__gpt-6.1-sol__medium__default__r70-7-go__r3", "codex__gpt-6.1-sol__medium__default__r70-5-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-5-ts-bun__r2", "codex__gpt-6.1-sol__medium__default__r70-5-ts-bun__r3", "codex__gpt-6.1-sol__medium__default__r70-6-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-6-ts-bun__r2", "codex__gpt-6.1-sol__medium__default__r70-6-ts-bun__r3", "codex__gpt-6.1-sol__medium__default__r70-7-ts-bun__r1", "codex__gpt-6.1-sol__medium__default__r70-7-ts-bun__r2", "codex__gpt-6.1-sol__medium__default__r70-7-ts-bun__r3"], "hypothesis_ids": ["H91", "H114"], "interpretation": "Rates are Rust 90.3%, Go 88.9%, and TS-Bun 83.3%; the raw combined totals remain Rust 18/20, Go 18/20, and TS-Bun 15/18. This selected extension and reused cohort do not establish a general language ranking.", "metric": "posthoc_equal_task_pass_rate", "numerator": null, "publication_status": "STATUS_ONLY", "reported_values": {"go": {"equal_task_rate_percent": 88.9, "raw_cells": 20, "raw_passes": 18, "task_fractions": {"r70-1-go": {"cells": 3, "passes": 2}, "r70-2-go": {"cells": 3, "passes": 3}, "r70-4-go-fe2": {"cells": 3, "passes": 2}, "r70-5-go": {"cells": 4, "passes": 4}, "r70-6-go": {"cells": 4, "passes": 4}, "r70-7-go": {"cells": 3, "passes": 3}}}, "rust": {"equal_task_rate_percent": 90.3, "raw_cells": 20, "raw_passes": 18, "task_fractions": {"r70-1-rust": {"cells": 3, "passes": 3}, "r70-2-rust": {"cells": 3, "passes": 3}, "r70-4-rust-fe2": {"cells": 3, "passes": 2}, "r70-5-rust": {"cells": 4, "passes": 4}, "r70-6-rust": {"cells": 4, "passes": 3}, "r70-7-rust": {"cells": 3, "passes": 3}}}, "ts-bun": {"equal_task_rate_percent": 83.3, "raw_cells": 18, "raw_passes": 15, "task_fractions": {"r70-1-ts-bun": {"cells": 3, "passes": 3}, "r70-2-ts-bun": {"cells": 3, "passes": 3}, "r70-4-ts-bun-fe2": {"cells": 3, "passes": 0}, "r70-5-ts-bun": {"cells": 3, "passes": 3}, "r70-6-ts-bun": {"cells": 3, "passes": 3}, "r70-7-ts-bun": {"cells": 3, "passes": 3}}}}, "round_id": "lang-sol-replication", "round_page": "rounds/lang-sol-replication/README.md", "schema_version": "1.0", "unit": "unweighted mean of six task-specific pass fractions"}
