Hallucination Detection on CT
94.33HDRUniHD
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| UniHDParam.=-, Backbone=MedGemma2026.06 | 94.33 | 56.84 | |
| UniHDParam.=-, Backbone=Grok-42026.06 | 87.86 | 52.56 | |
| UniHDParam.=-, Backbone=GPT-5.12026.06 | 82.01 | 58.16 | |
| UniHDParam.=-, Backbone=Gemini-32026.06 | 78.87 | 58.95 | |
| GPT-5.1-basedParam.=-, Backbone=Gemini-32026.06 | 71.13 | 64.71 | |
| GPT-5.1-basedParam.=-, Backbone=GPT-5.12026.06 | 71.07 | 63.43 | |
| GPT-5.1-basedParam.=-, Backbone=MedGemma2026.06 | 67.84 | 61.57 | |
| CounterVHDParam.=2B, Backbone=MedGemma2026.06 | 63.12 | 64.03 | |
| CounterVHDParam.=2B, Backbone=Gemini-32026.06 | 62.68 | 60.14 | |
| CounterVHDParam.=2B, Backbone=GPT-5.12026.06 | 61.15 | 68 | |
| CounterVHDParam.=2B, Backbone=Grok-42026.06 | 60 | 59.57 | |
| FaithscoreParam.=13B, Backbone=MedGemma2026.06 | 58.16 | 61.19 | |
| GPT-5.1-basedParam.=-, Backbone=Grok-42026.06 | 56.74 | 53.49 | |
| FaithscoreParam.=13B, Backbone=Grok-42026.06 | 55 | 65.81 | |
| FaithscoreParam.=13B, Backbone=Gemini-32026.06 | 52.11 | 64.35 | |
| FaithscoreParam.=13B, Backbone=GPT-5.12026.06 | 48.92 | 64.76 |