| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| OOD Evaluation (Squad, WQ, CWQ, MSQ, PopQA) | Thermometer | Squad Calibration Score5 | 11 | 3mo ago | |
| In-Domain Evaluation NQ, TQ, HQ, 2Wiki, Pararel | EliCal(1k) | Calibration Error (NQ)0.05 | 11 | 3mo ago | |
| ADVHOTPOT | P-E+EXPLCAL | AUC0.626 | 5 | 3mo ago | |
| HotpotQA (test) | LIMECAL | Accuracy65.7 | 5 | 3mo ago | |
| TriviaQA (test) | LIMECAL | Accuracy72 | 5 | 3mo ago | |
| SQuAD adversarial (test) | LIMECAL | Accuracy70.3 | 5 | 3mo ago |
