Mathematical Reasoning on Math Domain
66.45Avg AccuracyDVPO
Evaluation Results
| Method | Links | |
|---|---|---|
| DVPOTraining Domain=Math Domain2025.12 | 66.45 | |
| Dr.GRPOTraining Domain=Math Domain2025.12 | 61.64 | |
| Reinforce++Training Domain=Math Domain2025.12 | 61.35 | |
| Robust BellmanTraining Domain=Math Domain2025.12 | 60.56 | |
| GRPOTraining Domain=Math Domain2025.12 | 59.26 | |
| BaseTraining Domain=Math Domain2025.12 | 58.4 | |
| PPOTraining Domain=Math Domain2025.12 | 58.4 |