Hallucination Evaluation on HallusionBench GPT4-assisted (All)
49.94Accuracy (All)TGIF
Evaluation Results
| Method | Links | |
|---|---|---|
| TGIFBase Model=LLaVA-1.5, Params=7.0B2026.01 | 49.94 | |
| PerturboLLaVABase Model=LLaVA-1.5, Params=7.0B2026.01 | 47.6 | |
| OPERABase Model=LLaVA-1.5, Params=7.0B2026.01 | 47.1 | |
| LLaVA-1.5Params=13.0B2026.01 | 46.94 | |
| LLaVA-1.5Params=7.0B2026.01 | 46.9 | |
| VCDBase Model=LLaVA-1.5, Params=7.0B2026.01 | 46.9 | |
| InstructBLIPParams=8.2B2026.01 | 45.26 | |
| MiniGPT5Params=8.2B2026.01 | 40.3 | |
| Qwen-VLParams=9.6B2026.01 | 39.15 | |
| Open-FlamingoParams=9.0B2026.01 | 38.44 | |
| MiniGPT4Params=8.2B2026.01 | 35.78 |