Mathematical Reasoning on AMC (pass@1 statistics)
82.5Pass@1 MeanDAPO
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DAPOSampling strategy=avg@4, Number of repeats=42026.01 | 82.5 | 7.6 | |
| ReNCESampling strategy=avg@4, Number of repeats=42026.01 | 78.9 | 10.8 | |
| Semi-online DPOSampling strategy=avg@4, Number of repeats=42026.01 | 78 | 7.6 | |
| GRPOSampling strategy=avg@4, Number of repeats=42026.01 | 73.8 | 12 | |
| No trainingSampling strategy=avg@4, Number of repeats=42026.01 | 55.7 | 10.1 |