Math Reasoning on AIME '24 (Pass@1, Pass@k)
9.5Pass@1IMH
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| IMHModel=WeDLM-8B2026.04 | 9.5 | 34.4 | |
| EntropyModel=WeDLM-8B, Default decoding=true2026.04 | 9.3 | 31 | |
| ConfidenceModel=WeDLM-8B2026.04 | 7.9 | 32.3 | |
| MarginModel=WeDLM-8B2026.04 | 5.8 | 29.6 |