Image+Text-to-Text Hallucination Evaluation on MMHal-Bench
79BERT ScoreTIGER
Evaluation Results
| Method | Links | |
|---|---|---|
| TIGERBackbone=LLaVA-1.5-7B2026.05 | 79 | |
| BoN+VisPRMBackbone=LLaVA-1.5-7B2026.05 | 78.1 | |
| DeGFBackbone=LLaVA-1.5-7B2026.05 | 78 | |
| BoN+CLIPBackbone=LLaVA-1.5-7B2026.05 | 77.9 | |
| BoN+CycRewBackbone=LLaVA-1.5-7B2026.05 | 77.7 | |
| VCDBackbone=LLaVA-1.5-7B2026.05 | 77 | |
| VolcanoBackbone=LLaVA-1.5-7B2026.05 | 76.7 | |
| FrozenBackbone=LLaVA-1.5-7B2026.05 | 75.7 | |
| TIGERBackbone=Qwen2.5-Omni-7B2026.05 | 75.6 | |
| VolcanoBackbone=Qwen2.5-Omni-7B2026.05 | 74.2 | |
| VCDBackbone=Qwen2.5-Omni-7B2026.05 | 72 | |
| BoN+CycRewBackbone=Qwen2.5-Omni-7B2026.05 | 71.9 | |
| BoN+CLIPBackbone=Qwen2.5-Omni-7B2026.05 | 71.7 | |
| BoN+VisPRMBackbone=Qwen2.5-Omni-7B2026.05 | 71.4 | |
| WoodpeckerBackbone=Qwen2.5-Omni-7B2026.05 | 70.6 | |
| DeGFBackbone=Qwen2.5-Omni-7B2026.05 | 70.6 | |
| FrozenBackbone=Qwen2.5-Omni-7B2026.05 | 70.3 | |
| WoodpeckerBackbone=LLaVA-1.5-7B2026.05 | 67.2 |