Out-of-Distribution Detection on OBQA (In-distribution) and MMLU (Out-of-distribution)
91.7MPETN
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ETN2026.04 | 91.7 | 89.79 | 90.91 | |
| MAP_CE2026.04 | 89.83 | — | — | |
| IB-EDL2026.04 | 89.74 | 68.2 | 89.87 | |
| MCD2026.04 | 88.99 | 82.38 | — | |
| DeepEns2026.04 | 87.08 | 83.68 | — | |
| ETNBackbone=Gemma-2-9B2026.04 | 86.95 | 82.69 | 84.39 | |
| MAPCEBackbone=Gemma-2-9B2026.04 | 85.85 | — | — | |
| MCDBackbone=Gemma-2-9B2026.04 | 85.59 | 81.81 | — | |
| DMM2026.04 | 85.11 | 81.59 | 84.78 | |
| DeepEnsBackbone=Gemma-2-9B2026.04 | 83.04 | 78.82 | — | |
| DMMBackbone=Gemma-2-9B2026.04 | 81.48 | 80.53 | 81.51 | |
| MAPEDLBackbone=Gemma-2-9B2026.04 | 80.28 | 63.5 | 79.78 | |
| IB-EDLBackbone=Gemma-2-9B2026.04 | 79.99 | 66.58 | 74.28 | |
| MAP_EDL2026.04 | 79.86 | 70.16 | 79.95 | |
| LLBackbone=Gemma-2-9B2026.04 | 25.32 | 26.25 | — | |
| LL2026.04 | 24.74 | 24.63 | — |