Mathematical Reasoning on OlympiadBench (Acc, delta_Acc, Len, delta_percent_Len)
56.97AccAutoThink
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AutoThinkModel=DeepSeek-R1-Distill-Qwen-7B2025.08 | 56.97 | -0.15 | 4,817 | -45.19 | |
| OriginalModel=DeepSeek-R1-Distill-Qwen-7B2025.08 | 56.82 | — | 8,789 | — | |
| LCPOModel=DeepSeek-R1-Distill-Qwen-7B2025.08 | 56.08 | -0.74 | 4,222 | -51.96 | |
| AdaptThinkModel=DeepSeek-R1-Distill-Qwen-7B2025.08 | 55.93 | -0.89 | 6,594 | -24.97 | |
| LCPOModel=DeepSeek-R1-Distill-Qwen-1.5B2025.08 | 44.81 | 0.15 | 4,921 | -57.99 | |
| OriginalModel=DeepSeek-R1-Distill-Qwen-1.5B2025.08 | 44.66 | — | 11,715 | — | |
| AutoThinkModel=DeepSeek-R1-Distill-Qwen-1.5B2025.08 | 44.36 | -0.3 | 4,049 | -65.44 | |
| AdaptThinkModel=DeepSeek-R1-Distill-Qwen-1.5B2025.08 | 41.84 | -2.82 | 4,309 | -63.22 |