Mathematical Reasoning on GSM8K (Exact Match Accuracy)
95.9Exact Match AccuracyWORC+AC
Evaluation Results
| Method | Links | |
|---|---|---|
| WORC+AC2026.04 | 95.9 | |
| AoT2026.04 | 95 | |
| FoTn (sampling size)=82026.04 | 94 | |
| AFlow2026.04 | 93.5 | |
| AgentChain2026.04 | 92.4 | |
| CoT-SCn (sampling size)=52026.04 | 92 | |
| Self-Refine2026.04 | 91.7 | |
| CoT2026.04 | 90.9 | |
| Analogical Prompting2026.04 | 87.2 |