Mathematical Reasoning on AIME 24 (Maj@32, Avg@32)
66.7Majority Score @32PAEC
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PAECModel=Qwen3-4B-Base2026.06 | 66.7 | 48.3 | |
| GRPOModel=Qwen3-4B-Base2026.06 | 60 | 46.5 | |
| Global Entropy RegModel=Qwen3-4B-Base2026.06 | 60 | 46.9 | |
| BaseModel=Qwen3-4B-Base2026.06 | 56.7 | 46.5 | |
| GRPOModel=Llama3.1-8B2026.06 | 3.3 | 0.2 | |
| Global Entropy RegModel=Llama3.1-8B2026.06 | 3.3 | 0.2 | |
| PAECModel=Llama3.1-8B2026.06 | 3.3 | 0.1 | |
| BaseModel=Llama3.1-8B2026.06 | 0 | 0 |