Round r66

r66

Status · DESCRIPTIVEReviewed status is transcribed from rounds/STATUS.md. The original narrative and its qualifications remain below.

Source-reported narrativeRound README and results text are source material. Numeric values that are not tied to a claim ledger record remain source-reported.

Round source README · rounds/r66/README.md

Round date: HISTORICAL (before 2026-10-09) Publication badge: DESCRIPTIVE; KEPT FOR AUDIT Why not VALID: The historical inventory retains DESCRIPTIVE because of COHORT, DENOMINATOR, NO_PREREG. See the round evidence and limitations below. Recomputation status: OBSERVED SOURCE ONLY

Recomputation update: The mined raw records now provide per-cell outcome, manifest token-counter, and wall-time recomputation for the observed cohort. This does not reconstruct missing planned cells, a full intention-to-treat denominator, or model execution.

Status

DESCRIPTIVE

  • NO_PREREG — The README records no pre-registration or decision rule.
  • COHORT — Historical Sol-high and Luna-max control cohorts are not identified by public cell filters.
  • DENOMINATOR — The historical task comparisons cannot be reconstructed from the round files.

Required reproduction metadata

  • Kogen commit: not recorded in the cited round source.
  • Harness commit: not recorded in the cited round source as a Git commit; any adapter or runtime fingerprints are not Git commits.
  • Model and effort: a complete per-cell mapping is not recorded in the cited round source; values present in the source summary remain as stated above.
  • Historical execution command: not recorded in the cited round source.
  • Raw records: the cited sources do not provide a complete public round-specific request and grade bundle; public delivery and outcome exports are linked above where applicable.

Terminology: public round glossary.

STATUS: DESCRIPTIVE

Data completeness: 52.4% (20 deliveries; PASS WITH DECLARED GAPS). Measurement contract; declared gaps.

Status reason: The historical Sol-high and Luna-max cohorts in the prior headline are not identified by public cell filters; the historical counts and task comparisons are removed. The current medium cells remain available for descriptive reading.

Pre-registered: no

Headline-number note: The verdict figures are not re-derived here; the public results export is the outcome source.

Lifecycle: Historical capture. Snapshot 2026-10-05; earlier reports may use earlier cutoffs.

Question: How does direct Sol medium perform on four hard discriminators?

Venue: elx-07/erase-account/board-publish kogen-bench-us; elx-12 kogen-bench-eu.

Design: 4 tasks × 5 reps = 20; direct Codex, timeout 1740 s, zero retries. Historical Sol-high/Luna-max controls are referenced in prior material but their public cell cohorts are not identified.

Decision rule: none predeclared; descriptive n=5/task.

Arms: Codex gpt-6.1-sol medium. Historical Sol-high/Luna-max comparisons are withheld because their public cell cohorts are not identified.

Planned tasks in the surviving round note:

Verdict: The public cells export records 17/20 passes for direct Sol medium. Historical Sol-high/Luna-max figures and task-level comparisons are removed because their public cohorts and filters are not identified; no comparative claim is made.

Public delivery and outcome reconciliation

This captured-delivery summary uses the public run-record export, grouped by audit round, task ID, public host ID, and the exact exported arm label. Captured counts rows; graded counts rows marked graded=true; pass, fail, and other are the run-record outcome fields. Official outcome states are in cells.jsonl, whose classifications can differ; see the register source crosswalk. Planned n is shown only when source-reported; unknown allocations are not inferred.

Task ID Public host ID Arm label in export Planned n Captured Graded Pass Fail Other graded outcome
elx-07-cli-stats kogen-bench-us codex-solmed 5 5 5 5 0 0
elx-12-retry-api-deprecation kogen-bench-eu codex-solmed 5 5 5 5 0 0
elx-port-board-publish-unpublish-public-boundary kogen-bench-us codex-solmed 5 5 5 5 0 0
elx-port-erase-account kogen-bench-us codex-solmed 5 5 5 2 3 0

Recomputed from raw records

`recomputed.json`, mined cell records, and the token audit back observed official outcomes, per-arm pass counts/rates, manifest token counters, and wall medians. The source contains 20 rows (fail 3, pass 17); overall pass rate is 85.0% (17/20) among pass/fail grades. Per-arm detail and coverage counts are in the JSON.

Published-export outcome cross-check: 20/20 shared cell IDs match; 0/20 public-export rows have no mined source record. Exact IDs and results for unpaired rows are in recomputed.json. No per-cell outcome mismatches were found in the shared IDs. Token reconciliation under METHOD.md: 20/20 published metadata totals equal the manifest input + cached input + output sum; 0 differ (no mismatches). Reasoning remains a separate reported component and is not added to output. The recompute keeps both numbers visible. 0 unexplained single-response numeric errors were identified; multi-response rows retain the published full-cell total and the manifest counter separately because per-response usage records are excluded.

Captured run records

Standard-schema records assigned to this round in the indexed snapshot.

20 records · 126,749 bytesDownload r66.jsonl · View immutable source ↗

Reproduction inventory

Published analysis replay and full experiment replay are reported separately.

Replay type
unavailable
Script
No per-round analysis script is published.
Inputs
No identified public input bundle.
Full execution replay
Unavailable
Scope
No public per-round analysis script with a complete, identified input bundle is published. See the round README, MEASURED.md, and MISSING.md for recorded scope and gaps; historical execution inputs/toolchains/grading are not bundled.

View the complete per-round reproduction inventory · Repository reproduction guide