Unsupervised OOD detection on The Pile EDGAR Reports ID OOD (test)
68.09AUROCAP-OOD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AP-OODBase Model=Pythia-160M, ID Dataset=The Pile, OOD Type=Input OOD2026.02 | 68.09 | 91.47 | |
| Deep SVDDBase Model=Pythia-160M, ID Dataset=The Pile, OOD Type=Input OOD2026.02 | 64.06 | 88.94 | |
| KNNBase Model=Pythia-160M, ID Dataset=The Pile, OOD Type=Input OOD2026.02 | 59.41 | 93.26 | |
| MahalanobisBase Model=Pythia-160M, ID Dataset=The Pile, OOD Type=Input OOD2026.02 | 54.72 | 87.77 | |
| PerplexityBase Model=Pythia-160M, ID Dataset=The Pile, OOD Type=Input OOD2026.02 | 48.32 | 86.91 |