General Intelligence Evaluation on AGIEval G
72AccuracySA-SFT
Evaluation Results
| Method | Links | |
|---|---|---|
| SA-SFTBackbone=Qwen2.5-7B-I, Tuning Strategy=full, Prompting Setting=5-shot2026.01 | 72 | |
| BaseBackbone=Qwen2.5-7B-I, Tuning Strategy=full, Prompting Setting=5-shot2026.01 | 68.7 | |
| SA-SFTBackbone=LLaMA3-8B-I, Tuning Strategy=full, Prompting Setting=5-shot2026.01 | 29.8 | |
| SA-SFTBackbone=LLaMA2-13B-C, Tuning Strategy=LoRA, Prompting Setting=5-shot2026.01 | 26.4 |