Knowledge Intensive on Gaokao Geography
79.8AccuracyMemAPO
Evaluation Results
| Method | Links | |
|---|---|---|
| MemAPOBackbone=Qwen3-8B2026.03 | 79.8 | |
| RaRBackbone=Qwen3-8B2026.03 | 78.8 | |
| CoTBackbone=Qwen3-8B2026.03 | 76.8 | |
| TextGradBackbone=Qwen3-8B2026.03 | 76.8 | |
| ProTeGiBackbone=Qwen3-8B2026.03 | 75.8 | |
| TextGradBackbone=GPT-4o-mini2026.03 | 73.7 | |
| MemAPOBackbone=GPT-4o-mini2026.03 | 73.7 | |
| IOBackbone=Qwen3-8B2026.03 | 72.7 | |
| OPROBackbone=Qwen3-8B2026.03 | 72.7 | |
| SPOBackbone=GPT-4o-mini2026.03 | 72.7 | |
| SPOBackbone=Qwen3-8B2026.03 | 71.7 | |
| IOBackbone=GPT-4o-mini2026.03 | 71.7 | |
| OPROBackbone=GPT-4o-mini2026.03 | 71.7 | |
| Step-BackBackbone=Qwen3-8B2026.03 | 70.7 | |
| PromptBreederBackbone=Qwen3-8B2026.03 | 70.7 | |
| CoTBackbone=GPT-4o-mini2026.03 | 70.7 | |
| ProTeGiBackbone=GPT-4o-mini2026.03 | 68.7 | |
| RaRBackbone=GPT-4o-mini2026.03 | 67.7 | |
| PromptBreederBackbone=GPT-4o-mini2026.03 | 66.7 | |
| Step-BackBackbone=GPT-4o-mini2026.03 | 62.6 |