Question Answering on SciQ (ECE', Brier', AUAC)
0ECE' GainNone
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| NoneCalibration Method=None2026.04 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | |
| S-BCalibration Method=Self-Binary2026.04 | 0.748 | 0 | 0.567 | 0 | 0.777 | 0 | -0.61 | 0 | |
| HS-QABCalibration Method=HS-QAB2026.04 | 0.776 | 0 | 0.916 | 0 | 0.693 | 0 | 0.024 | 1 | |
| S-QABCalibration Method=S-QAB2026.04 | 0.837 | 0 | 1.027 | 1 | 0.693 | 0 | 0.015 | 0 | |
| BCalibration Method=Binary2026.04 | 0.901 | 0 | 0.648 | 0 | 0.937 | 0 | -0.039 | 0 | |
| QABCalibration Method=QAB2026.04 | 1.09 | 0 | 0.99 | 1 | 0.936 | 0 | 0.068 | 0 | |
| GIRBCalibration Method=Group Isotonic Regression Binning2026.04 | 1.164 | 0 | 1.006 | 0 | 1.009 | 4 | 0.127 | 3 | |
| SCalibration Method=Self2026.04 | 1.588 | 4 | 1.085 | 2 | 0.995 | 0 | -0.66 | 0 |