By: ValentinHirschi <valentin.hirschi@gmail.com>
Add native tangent IBP sources and checked priority-rule experiments Regenerate exact ordinary-source combinations with guard and uniform-descent checks. Preserve candidate-owner fallback rules and terminals. Record successful native/cold-load controls and the negative whole-cohort performance result without promoting the new priority policy. Co-authored-by: Codex <noreply@openai.com>
| Time to Start | Worker time | Duration | Time to finish | Idle | |
|---|---|---|---|---|---|
| Config | - | - | - | - | 470h30m26s |
| Eval | - | - | - | - | - |
| Build | - | - | - | - | - |
| Suite | - | - | - | - | 470h30m26s |