Mathematical Reasoning on AMC 23 (Pass@32)
95.6Pass@32OPEFO
Evaluation Results
| Method | Links | |
|---|---|---|
| OPEFOBackbone=Qwen2.5-Math-7B2026.05 | 95.6 | |
| Clip-CovBackbone=Qwen2.5-Math-7B2026.05 | 95.3 | |
| KL-CovBackbone=Qwen2.5-Math-7B2026.05 | 94.9 | |
| GRPO (Strict)Backbone=Qwen2.5-Math-7B2026.05 | 94.8 | |
| Clip-higherBackbone=Qwen2.5-Math-7B2026.05 | 94.6 | |
| GRPOBackbone=Qwen2.5-Math-7B2026.05 | 93.7 | |
| Entropy-RegBackbone=Qwen2.5-Math-7B2026.05 | 90.8 |