Math Word Problems on SVAMP (Source/Unseen Library Comparison)
93.64Accuracy (Source Lib.)SIGMA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SIGMABase Model=GPT-4o-mini2026.06 | 93.64 | 93.43 | 0.21 | |
| CARDBase Model=GPT-4o-mini2026.06 | 91.65 | 90.79 | 0.86 | |
| G-DesignerBase Model=GPT-4o-mini2026.06 | 90.32 | 89.17 | 1.15 | |
| GPTSwarmBase Model=GPT-4o-mini2026.06 | 90.12 | 88.14 | 1.98 | |
| LLM-DebateBase Model=GPT-4o-mini2026.06 | 87.43 | 85.46 | 1.97 |