Performance Shift Prediction on MMLU-Pro
0.8R-squaredSTS_act
Evaluation Results
| Method | Links | |
|---|---|---|
| STS_actModel=Qwen2.5-7B, SFT Dataset=LIMO2026.03 | 0.8 | |
| STS_ICLModel=LLaMA3-8B, SFT Dataset=LIMO2026.03 | 0.66 | |
| STS_ICLModel=Qwen2.5-7B, SFT Dataset=LIMO2026.03 | 0.61 | |
| STS_ICLModel=Gemma2-9B, SFT Dataset=LIMO2026.03 | 0.6 | |
| STS_actModel=LLaMA3-8B, SFT Dataset=LIMO2026.03 | 0.5 | |
| STS_actModel=Gemma2-9B, SFT Dataset=LIMO2026.03 | 0.36 |