Mathematical Reasoning on AIME OOD 24 (Pass@1, Tokens)
0.5Pass@1 AccuracyBase
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BaseBackbone=R1-Distill-Qwen-7B, Evaluation Setting=zero-shot2025.05 | 0.5 | 8,674 | |
| DRPBackbone=R1-Distill-Qwen-7B, Evaluation Setting=zero-shot2025.05 | 0.5 | 4,966 | |
| Cot ValveBackbone=R1-Distill-Qwen-7B, Evaluation Setting=zero-shot2025.05 | 0.4333 | 6,315 | |
| DRPBackbone=R1-Distill-Qwen-1.5B, Evaluation Setting=zero-shot2025.05 | 0.3333 | 6,135 | |
| TALEBackbone=R1-Distill-Qwen-7B, Evaluation Setting=zero-shot2025.05 | 0.3333 | 8,602 | |
| ThinkpruneBackbone=R1-Distill-Qwen-1.5B, Evaluation Setting=zero-shot2025.05 | 0.3 | 5,745 | |
| Cot ValveBackbone=R1-Distill-Qwen-1.5B, Evaluation Setting=zero-shot2025.05 | 0.2333 | 5,601 | |
| BaseBackbone=R1-Distill-Qwen-1.5B, Evaluation Setting=zero-shot2025.05 | 0.2 | 10,484 | |
| TALEBackbone=R1-Distill-Qwen-1.5B, Evaluation Setting=zero-shot2025.05 | 0.2 | 8,915 |