Mathematical Reasoning on AMC (Avg@32, Pass@32)
63.8Avg@32EntroAdv
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EntroAdvBackbone=Qwen2.5-Math-7B, Method Category=Policy Entropy Method2026.04 | 63.8 | 89.4 | |
| ForkingBackbone=Qwen2.5-Math-7B, Method Category=Policy Entropy Method2026.04 | 63.7 | 86.6 | |
| EAPOBackbone=Qwen2.5-Math-7B, Method Category=Policy Entropy Method2026.04 | 62.1 | 88.8 | |
| EntroRegBackbone=Qwen2.5-Math-7B, Method Category=Policy Entropy Method2026.04 | 61.1 | 84.9 | |
| DAPOBackbone=Qwen2.5-Math-7B, Method Category=Classic RLVR Method2026.04 | 61 | 84.8 | |
| KL-CovBackbone=Qwen2.5-Math-7B, Method Category=Policy Entropy Method2026.04 | 59.2 | 85.9 | |
| GRPOBackbone=Qwen2.5-Math-7B, Method Category=Classic RLVR Method2026.04 | 58.6 | 80.3 | |
| W-REINFBackbone=Qwen2.5-Math-7B, Method Category=Classic RLVR Method2026.04 | 58.1 | 87.1 | |
| PRIMEBackbone=Qwen2.5-Math-7B, Method Category=Classic RLVR Method2026.04 | 42.4 | 90.5 | |
| Base-7BBackbone=Qwen2.5-Math-7B, Method Category=Classic RLVR Method2026.04 | 32.7 | 85.2 |