Hallucination Detection on HealthBench (test)
96.48AUCLinear Probe
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Linear Probe2025.12 | 96.48 | 73.5 | 96.48 | 39.62 | 37.98 | |
| MLP Probe2025.12 | 95.49 | 76.95 | 95.49 | 32.31 | 49.04 | |
| Entropy2025.12 | 47.15 | 9.48 | — | — | — | |
| Perplexity2025.12 | 46.05 | 8.12 | — | — | — |