Expert-Iteration RLVR on MedQA, HEAD-QA, ARC-C, and CaseHOLD
4Pathwise Clean Score
CSA
Updated 13d ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.05 | 4 | 4 | 4 | 52 | |
2026.05 | 4 | 0 | 0 | 2.8 | |
2026.05 | 4 | 1 | 1 | 14.2 | |
2026.05 | 3 | 1 | 1 | 25.5 | |
2026.05 | 2 | 4 | 2 | 66.1 | |
2026.05 | 1 | 4 | 1 | 79.8 | |
2026.05 | 1 | 4 | 1 | 90.7 | |
2026.05 | 0 | 4 | 0 | 90.3 | |
2026.05 | 0 | 4 | 0 | 87.2 | |
2026.05 | 0 | 4 | 0 | 100 |