Hallucination Detection on MMMU
66.7AUROCSCoOP
Evaluation results
| Method | Links | |
|---|---|---|
| SCoOP#VLMs=4-VLM2026.03 | 66.7 | |
| SCoOP#VLMs=3-VLM2026.03 | 64.6 | |
| SCoOP#VLMs=2-VLM2026.03 | 62.7 | |
| Naive Selection#VLMs=2-VLM2026.03 | 61.5 | |
| Naive Selection#VLMs=3-VLM2026.03 | 60.2 | |
| Majority Voting#VLMs=4-VLM2026.03 | 58.3 | |
| Naive Selection#VLMs=4-VLM2026.03 | 57.6 | |
| Majority Voting#VLMs=3-VLM2026.03 | 55.4 | |
| Majority Voting#VLMs=2-VLM2026.03 | 53 |