Mathematical Reasoning on MATH500 (mean accuracy)
56Mean Accuracy (MATH500)Cliff-DPO (uncertain + sampled-off)
Evaluation results
| Method | Links | |
|---|---|---|
| Cliff-DPO (uncertain + sampled-off)evaluation_protocol=avg@64, Updated tokens=32,9162026.06 | 56 | |
| Qwen3-0.6B + cDPOevaluation_protocol=avg@64, Updated tokens=5,829,0522026.06 | 54.9 | |
| Cliff-DPO (all)evaluation_protocol=avg@64, Updated tokens=38,4542026.06 | 53.8 | |
| Cliff-DPO (uncertain)evaluation_protocol=avg@64, Updated tokens=18,1222026.06 | 53.3 | |
| Cliff-DPO (sampled-off)evaluation_protocol=avg@64, Updated tokens=14,7942026.06 | 52.6 | |
| Qwen3-0.6Bevaluation_protocol=avg@642026.06 | 51.6 | |
| Cliff-DPO (deterministic)evaluation_protocol=avg@64, Updated tokens=5,5382026.06 | 51.5 | |
| Qwen3-0.6B + DPOevaluation_protocol=avg@64, Updated tokens=2,862,8452026.06 | 51 |