Everything behind the paper: the data, the code that produced every table, and the ledger of every claim considered, including the ones that failed checking.
| file | size | what it is |
|---|---|---|
| 2026-08-01_192240_baseline-v3.json | 32.9 KB | structured output kept exactly as recorded |
| 2026-08-01_195324_v5-final.json | 33.6 KB | structured output kept exactly as recorded |
| 2026-08-01_201419_v6-final.json | 34.1 KB | structured output kept exactly as recorded |
| 2026-08-01_201640_v7-final.json | 32.6 KB | structured output kept exactly as recorded |
| 2026-08-01_203012_v8-baseline.json | 33.4 KB | structured output kept exactly as recorded |
| 2026-08-02_014852_v9-post-testing.json | 33.3 KB | structured output kept exactly as recorded |
| 2026-08-02_020605_v10-stable-baseline.json | 36.5 KB | structured output kept exactly as recorded |
| PLAN.md | 7.7 KB | a note published with the paper |
| PROGRESS.md | 1.5 KB | what was done, what is open, and what was found wrong |
| analyse.py | 4.2 KB | the code that produced or checked the numbers |
| assemble.py | 3.0 KB | the code that produced or checked the numbers |
| benchmark-tasks.json | 18.5 KB | structured output kept exactly as recorded |
| claim-ledger.csv | 10.4 KB | every claim in the paper, its status and its source |
| extract.py | 3.2 KB | the code that produced or checked the numbers |
| licence.txt | 0.5 KB | the licence for this package |
| negative-control-results.json | 4.9 KB | structured output kept exactly as recorded |
| negative-control.py | 4.4 KB | the code that produced or checked the numbers |
| timeline.py | 2.5 KB | the code that produced or checked the numbers |
| versions-by-task.csv | 40.0 KB | a table, recomputable by anyone holding this file |
| runs/2026-08-01_192240_baseline-v3.json | 32.9 KB | structured output kept exactly as recorded |
| runs/2026-08-01_195324_v5-final.json | 33.6 KB | structured output kept exactly as recorded |
| runs/2026-08-01_201419_v6-final.json | 34.1 KB | structured output kept exactly as recorded |
| runs/2026-08-01_201640_v7-final.json | 32.6 KB | structured output kept exactly as recorded |
| runs/2026-08-01_203012_v8-baseline.json | 33.4 KB | structured output kept exactly as recorded |
| runs/2026-08-02_014852_v9-post-testing.json | 33.3 KB | structured output kept exactly as recorded |
| runs/2026-08-02_020605_v10-stable-baseline.json | 36.5 KB | structured output kept exactly as recorded |
Every figure in the paper can be recomputed from these files. If one of them disagrees with the paper, the file is right and the paper is wrong, and it will be corrected in the change log.