Mathematical Reasoning on MATH (Accuracy, Throughput)
76.28AccuracyQwen2.5 7B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen2.5 7BGeneration mode=Autoregressive2026.06 | 76.28 | 46.5 | |
| PoE-BridgeImportance sampling=true2026.06 | 74.42 | 94.94 | |
| PoE-BridgeImportance sampling=false2026.06 | 73.86 | 99.49 | |
| Dream 7BSampling strategy=Entropy, Tokens per step=22026.06 | 27.91 | 29.13 | |
| Dream 7BSampling strategy=Random, Tokens per step=22026.06 | 14.57 | 30.59 | |
| Dream 7BSampling strategy=Entropy, Tokens per step=42026.06 | 11.73 | 47.81 |