Mathematical Reasoning on GSM8K (VAcc, RAcc, Delta Acc, TFR)
92.34Verification Accuracy (VAcc)MASTER
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MASTERTarget model=Qwen2.5-Math-7B-Instruct, Prompting=zero-shot2025.06 | 92.34 | 48.63 | 43.71 | 47.33 | |
| MASTERTarget model=DeepSeek-R1-Distill-LLaMA-8B, Prompting=zero-shot2025.06 | 83.09 | 31.85 | 51.24 | 61.67 | |
| MASTERTarget model=LLaMA-3-8B-Instruct, Prompting=zero-shot2025.06 | 76.42 | 27.51 | 48.91 | 64 | |
| MASTERTarget model=Gemma-2-2B-IT, Prompting=zero-shot2025.06 | 63.15 | 13.68 | 49.47 | 78.33 | |
| MASTERTarget model=LLaMA-3.2-1B-Instruct, Prompting=zero-shot2025.06 | 45.26 | 2.71 | 42.55 | 94 |