Hallucination Mitigation on MME Existence, Count, Position, Color
195Existence ScoreVAF
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| VAFModel=LLaVA-v1.5-7B2025.03 | 195 | 158.33 | 128.33 | 155 | 636.67 | |
| VAFModel=LLaVA-v1.5-13B2025.03 | 195 | 160 | 136.67 | 170 | 661.67 | |
| RegularModel=LLaVA-v1.5-7B2025.03 | 185 | 146.67 | 128.33 | 150 | 610 | |
| VCDModel=LLaVA-v1.5-7B2025.03 | 185 | 141.33 | 128.33 | 153 | 607.66 | |
| ICDModel=LLaVA-v1.5-7B2025.03 | 185 | 148.33 | 126.66 | 148.33 | 608.32 | |
| RegularModel=LLaVA-v1.5-13B2025.03 | 185 | 155 | 133.33 | 165 | 638.33 | |
| VCDModel=LLaVA-v1.5-13B2025.03 | 185 | 155 | 130 | 168.33 | 638.33 | |
| ICDModel=LLaVA-v1.5-13B2025.03 | 183.33 | 153.33 | 131.67 | 165 | 633.33 | |
| VAFModel=Qwen-VL-7B2025.03 | 165 | 155 | 133.33 | 175 | 628.33 | |
| RegularModel=Qwen-VL-7B2025.03 | 158.33 | 150 | 128.33 | 170 | 606.66 | |
| VCDModel=Qwen-VL-7B2025.03 | 158.33 | 150 | 133.33 | 175 | 616.66 | |
| ICDModel=Qwen-VL-7B2025.03 | 128.33 | 151.67 | 128.33 | 170 | 578.33 |