Hallucination Detection on SLAKE
71.9AUCCEBaG
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| CEBaGMLLM Backbone=MedGemma-4b-it2026.03 | 71.9 | 75.9 | |
| CEBaGMLLM Backbone=HuatuoGPT-Vision-7B2026.03 | 67 | 64.2 | |
| CEBaGMLLM Backbone=LLaVA-Med-v1.5-Mistral-7B2026.03 | 63.8 | 52.3 | |
| VASEMLLM Backbone=MedGemma-4b-it2026.03 | 62.6 | 70.1 | |
| RadFlagMLLM Backbone=HuatuoGPT-Vision-7B2026.03 | 60.6 | 54.2 | |
| VASEMLLM Backbone=HuatuoGPT-Vision-7B2026.03 | 59.9 | 58 | |
| AvgProbMLLM Backbone=LLaVA-Med-v1.5-Mistral-7B2026.03 | 59.9 | 50 | |
| VASEMLLM Backbone=MedGemma-1.5-4b-it2026.03 | 59.5 | 52.9 | |
| RadFlagMLLM Backbone=LLaVA-Med-v1.5-Mistral-7B2026.03 | 58.7 | 50.7 | |
| SEMLLM Backbone=LLaVA-Med-v1.5-Mistral-7B2026.03 | 57.9 | 50.2 | |
| SEMLLM Backbone=MedGemma-1.5-4b-it2026.03 | 57.7 | 51.9 | |
| SEMLLM Backbone=HuatuoGPT-Vision-7B2026.03 | 57.7 | 54.9 | |
| CEBaGMLLM Backbone=MedGemma-1.5-4b-it2026.03 | 57.5 | 52 | |
| AvgProbMLLM Backbone=MedGemma-1.5-4b-it2026.03 | 57.4 | 52 | |
| VASEMLLM Backbone=LLaVA-Med-v1.5-Mistral-7B2026.03 | 57.4 | 50.8 | |
| RadFlagMLLM Backbone=MedGemma-1.5-4b-it2026.03 | 57.2 | 53 | |
| RadFlagMLLM Backbone=MedGemma-4b-it2026.03 | 57 | 66.2 | |
| SEMLLM Backbone=MedGemma-4b-it2026.03 | 54.9 | 64.2 | |
| AvgProbMLLM Backbone=MedGemma-4b-it2026.03 | 35.3 | 52.9 | |
| AvgProbMLLM Backbone=HuatuoGPT-Vision-7B2026.03 | 30.5 | 38.4 |