Mathematical Reasoning on GSM8K (pass@K)
97.77pass@KRECAP
Evaluation Results
| Method | Links | |
|---|---|---|
| RECAPBackbone=DSQwen-14B2025.10 | 97.77 | |
| DAPOBackbone=DSQwen-14B2025.10 | 97.19 | |
| SafeChainBackbone=DSQwen-14B2025.10 | 96.44 | |
| SFTBackbone=DSQwen-14B2025.10 | 95.9 | |
| STARBackbone=DSQwen-14B2025.10 | 95.86 | |
| OriginalBackbone=DSQwen-14B2025.10 | 95.2 | |
| RECAPBackbone=DSLlama-8B2025.10 | 93.72 | |
| DAPOBackbone=DSLlama-8B2025.10 | 93.71 | |
| SafeChainBackbone=DSLlama-8B2025.10 | 91.32 | |
| SFTBackbone=DSLlama-8B2025.10 | 91.32 | |
| STARBackbone=DSLlama-8B2025.10 | 90.74 | |
| OriginalBackbone=DSLlama-8B2025.10 | 90.32 |