Near-OOD Detection on CLINC150
86.61AUROCGemma2-2b LR baseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Gemma2-2b LR baselineBackbone=Gemma2-2b, Detection Backend=Logistic Regression baseline, Evaluation Protocol=ID-only training2026.05 | 86.61 | 64.49 | 56.07 | |
| Gemma2-9b LR baselineBackbone=Gemma2-9b, Detection Backend=Logistic Regression baseline, Evaluation Protocol=ID-only training2026.05 | 84.94 | 62.96 | 64.4 | |
| Gemma2-9b + 16k SAE + MarkovBackbone=Gemma2-9b, SAE Configuration=16k, Detection Backend=Markov, Evaluation Protocol=ID-only training2026.05 | 83.64 | 72.78 | 48.25 | |
| Gemma2-2b + 16k SAE + MarkovBackbone=Gemma2-2b, SAE Configuration=16k, Detection Backend=Markov, Evaluation Protocol=ID-only training2026.05 | 73.87 | 1.27 | 83.33 |