Mathematical Reasoning on ProofNet (test)
95.6AccuracyReProbe
Evaluation Results
| Method | Links | |
|---|---|---|
| ReProbe# Sample=32K, Input Features=Hidden States, Annotation Source=DeepSeek-anno, Decoding Method=Beam search2025.11 | 95.6 | |
| Qwen2.5-Math-7B-PRM800k# Sample=263K, Decoding Method=Beam search2025.11 | 95.2 | |
| ReProbe# Sample=32K, Input Features=Attn+Logit, Annotation Source=Self-anno, Decoding Method=Beam search2025.11 | 95.1 | |
| Qwen2.5-Math-PRM-7B# Sample=860K, Decoding Method=Beam search2025.11 | 93.6 | |
| ReProbe# Sample=32K, Input Features=Hidden States, Annotation Source=Self-anno, Decoding Method=Beam search2025.11 | 90.6 | |
| ReProbe# Sample=32K, Input Features=Attn+Logit, Annotation Source=DeepSeek-anno, Decoding Method=Beam search2025.11 | 79.1 |