Object Hallucination Evaluation on GQA Popular split
86.07AccuracyMESA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MESAModel=Qwen-VL2026.04 | 86.07 | 84.89 | |
| VAFModel=Qwen-VL2026.04 | 85.43 | 81.35 | |
| ICTModel=Qwen-VL2026.04 | 85.31 | 84.15 | |
| CAIBackbone=Qwen-VL2026.06 | 84.97 | 84.95 | |
| VCDBackbone=Qwen-VL2026.06 | 84.33 | 84.82 | |
| VTIModel=Qwen-VL2026.04 | 83.67 | 81.31 | |
| CAIBackbone=LLaVA-1.52026.06 | 83.43 | 84.39 | |
| MESAModel=LLaVA-v1.52026.04 | 82.93 | 83.91 | |
| ICTModel=LLaVA-v1.52026.04 | 82.86 | 83.7 | |
| ONLYBackbone=Qwen-VL2026.06 | 82.8 | 82.72 | |
| PAIBackbone=Qwen-VL2026.06 | 82.67 | 81.89 | |
| VCDModel=Qwen-VL2026.04 | 81.83 | 82.23 | |
| VTIModel=LLaVA-v1.52026.04 | 81.73 | 82.39 | |
| CAIBackbone=InstructBLIP2026.06 | 81.2 | 82.05 | |
| VAFModel=LLaVA-v1.52026.04 | 81.16 | 82 | |
| ICDModel=Qwen-VL2026.04 | 80.8 | 77.89 | |
| RegularBackbone=Qwen-VL2026.06 | 80.77 | 80.15 | |
| VCDModel=LLaVA-v1.52026.04 | 80.73 | 82.24 | |
| ICDModel=LLaVA-v1.52026.04 | 78.56 | 78.79 | |
| NulluModel=LLaVA-v1.52026.04 | 77.63 | 80.13 | |
| VCDBackbone=InstructBLIP2026.06 | 77.57 | 79.4 | |
| ONLYBackbone=LLaVA-1.52026.06 | 75.93 | 79.62 | |
| ONLYBackbone=InstructBLIP2026.06 | 74.8 | 78.41 | |
| VanillaModel=Qwen-VL2026.04 | 74.76 | 66.66 | |
| NulluModel=Qwen-VL2026.04 | 74.71 | 66.54 | |
| VCDBackbone=LLaVA-1.52026.06 | 74.37 | 78.97 | |
| PAIBackbone=InstructBLIP2026.06 | 73.77 | 77.34 | |
| VanillaModel=LLaVA-v1.52026.04 | 73.3 | 77.93 | |
| PAIBackbone=LLaVA-1.52026.06 | 72.73 | 77.6 | |
| RegularBackbone=InstructBLIP2026.06 | 72.27 | 75.97 | |
| RegularBackbone=LLaVA-1.52026.06 | 71.93 | 76.88 |