Language Understanding on MMLU-Pro (Brier score)
19.26Brier ScoreCAGE-CAL
Evaluation Results
| Method | Links | |
|---|---|---|
| CAGE-CALCalibration category=Trained calibrators2026.05 | 19.26 | |
| DiscoUQ-LLMCalibration category=Trained calibrators2026.05 | 24.02 | |
| + Scaling-bin.Calibration category=Post-hoc plurality calibrators2026.05 | 24.21 | |
| + PlattCalibration category=Post-hoc plurality calibrators2026.05 | 24.3 | |
| + IsotonicCalibration category=Post-hoc plurality calibrators2026.05 | 24.76 | |
| Plurality shareCalibration category=Post-hoc plurality calibrators2026.05 | 25.01 | |
| Scalar + GBTCalibration category=Trained calibrators2026.05 | 25.96 | |
| GraphCalCalibration category=Trained calibrators2026.05 | 26.28 | |
| Collab. Cal.Calibration category=LLM-elicited confidence estimators2026.05 | 33.71 | |
| LLM-Cal (+topo)Calibration category=LLM-elicited confidence estimators2026.05 | 34.6 | |
| LLM-Cal (no topo)Calibration category=LLM-elicited confidence estimators2026.05 | 35.46 |