Hallucination Detection on WNMCQ1 (AUROC)
91.2AUROCUQ_ICL
Evaluation Results
| Method | Links | |
|---|---|---|
| UQ_ICLModel=LLaMA2-13B2026.04 | 91.2 | |
| Semantic EntropyModel=LLaMA2-13B2026.04 | 90.4 | |
| Self-FVModel=LLaMA2-7B2026.04 | 89.93 | |
| Function VectorModel=LLaMA2-13B2026.04 | 89.9 | |
| Function VectorModel=LLaMA2-7B2026.04 | 89.89 | |
| Self-FVModel=LLaMA2-13B2026.04 | 88.9 | |
| UQ_ICLModel=LLaMA2-7B2026.04 | 88.4 | |
| MaxProbModel=LLaMA2-7B2026.04 | 87.5 | |
| Total EntropyModel=LLaMA2-13B2026.04 | 87.3 | |
| Total EntropyModel=LLaMA2-7B2026.04 | 87.2 | |
| Semantic EntropyModel=LLaMA2-7B2026.04 | 85.9 | |
| Lookback LensModel=LLaMA2-7B2026.04 | 57.2 |