Mathematical Calculation on AQuA-RAT
83.1Accuracy (AQuA-RAT)MemAPO
Evaluation Results
| Method | Links | |
|---|---|---|
| MemAPOBackbone=Qwen3-8B2026.03 | 83.1 | |
| CoTBackbone=Qwen3-8B2026.03 | 82.5 | |
| OPROBackbone=Qwen3-8B2026.03 | 82.5 | |
| MemAPOBackbone=GPT-4o-mini2026.03 | 82.5 | |
| Step-BackBackbone=Qwen3-8B2026.03 | 81.8 | |
| ProTeGiBackbone=Qwen3-8B2026.03 | 81.2 | |
| RaRBackbone=Qwen3-8B2026.03 | 80.5 | |
| TextGradBackbone=Qwen3-8B2026.03 | 80.5 | |
| IOBackbone=Qwen3-8B2026.03 | 79.9 | |
| TextGradBackbone=GPT-4o-mini2026.03 | 79.9 | |
| OPROBackbone=GPT-4o-mini2026.03 | 79.2 | |
| PromptBreederBackbone=GPT-4o-mini2026.03 | 78.6 | |
| SPOBackbone=GPT-4o-mini2026.03 | 77.9 | |
| ProTeGiBackbone=GPT-4o-mini2026.03 | 76 | |
| PromptBreederBackbone=Qwen3-8B2026.03 | 74 | |
| SPOBackbone=Qwen3-8B2026.03 | 72.7 | |
| CoTBackbone=GPT-4o-mini2026.03 | 69.5 | |
| RaRBackbone=GPT-4o-mini2026.03 | 66.9 | |
| IOBackbone=GPT-4o-mini2026.03 | 61.7 | |
| Step-BackBackbone=GPT-4o-mini2026.03 | 59.7 |