Visual Hallucination Mitigation on GPT-4 assisted benchmark
46.7SHRGreedy
Evaluation Results
| Method | Links | |
|---|---|---|
| GreedyBackbone=MiniGPT-42025.01 | 46.7 | |
| VCDBackbone=MiniGPT-42025.01 | 46 | |
| OPERABackbone=MiniGPT-42025.01 | 45.9 | |
| HALCBackbone=MiniGPT-42025.01 | 45.8 | |
| VASparseBackbone=MiniGPT-42025.01 | 45.2 | |
| GreedyBackbone=mPLUG-Owl22025.01 | 42.3 | |
| VCDBackbone=mPLUG-Owl22025.01 | 41.9 | |
| OPERABackbone=mPLUG-Owl22025.01 | 41.7 | |
| HALCBackbone=mPLUG-Owl22025.01 | 41.7 | |
| VASparseBackbone=mPLUG-Owl22025.01 | 41.1 | |
| GreedyBackbone=LLaVA-1.52025.01 | 36.3 | |
| VCDBackbone=LLaVA-1.52025.01 | 34.6 | |
| OPERABackbone=LLaVA-1.52025.01 | 34.2 | |
| HALCBackbone=LLaVA-1.52025.01 | 33.9 | |
| VASparseBackbone=LLaVA-1.52025.01 | 33.5 |