Question Answering on MMLU (F1, AUARC)
4.37F1MAD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MADConfidence Estimator=Sequence Probability, Calibration=Beta2026.06 | 4.37 | 7.63 | |
| MADConfidence Estimator=Self-Report, Calibration=Beta2026.06 | 4.37 | 10.05 | |
| WSVConfidence Estimator=Self-Report, Calibration=Beta2026.06 | 4.23 | 9.89 | |
| WSVConfidence Estimator=Sequence Probability, Calibration=Beta2026.06 | 3.74 | 8.95 | |
| CGAConfidence Estimator=Self-Report, Calibration=Beta2026.06 | 3.58 | 3.51 | |
| HIDConfidence Estimator=Self-Report, Calibration=Beta2026.06 | 3.36 | 4.26 | |
| MAD-DConfidence Estimator=Sequence Probability, Calibration=Beta2026.06 | 3.14 | 2.48 | |
| MAD-DConfidence Estimator=Self-Report, Calibration=Beta2026.06 | 3.14 | 6.14 | |
| HIDConfidence Estimator=Sequence Probability, Calibration=Beta2026.06 | 2.54 | 5.76 | |
| CGAConfidence Estimator=Sequence Probability, Calibration=Beta2026.06 | 2.35 | 4.73 |