Error detection on CEBaB (test)
0.81Rho(U_epi, U_ale)Semantic Entropy
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Semantic Entropy2026.02 | 0.81 | 0.22 | 0.68 | |
| Deep Ensembles2026.02 | 0.79 | 0.24 | 0.7 | |
| MC Dropout2026.02 | 0.75 | 0.19 | 0.65 | |
| Credal CBM2026.02 | 0.08 | 0.74 | 0.78 |
Loading the SOTA2 catalog…