Mathematical Reasoning on MATH-500 n=200 (pilot)
69.5Accuracy (%)Town@1024
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Town@1024Model=Qwen3-8B, Decoding Strategy=greedy decoding2026.05 | 69.5 | 877 | |
| Mrsd (3-round)Model=Qwen3-8B, Decoding Strategy=greedy decoding2026.05 | 61 | 1,823 | |
| Nothink@B1Model=Qwen3-8B, Decoding Strategy=greedy decoding2026.05 | 47.5 | 418 | |
| Think@BthinkModel=Qwen3-8B, Decoding Strategy=greedy decoding2026.05 | 19.5 | 1,024 |