Hallucination Evaluation on MME Color
170Score (Chinese)Baseline
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| BaselineBackbone=Qwen-VL-Chat2025.06 | 170 | 180 | 150 | 153.3 | — | — | 103.3 | 165 | 148.3 | |
| CLAIMBackbone=Qwen-VL-Chat2025.06 | 165 | 170 | 158.3 | 90 | — | — | 165 | — | 149.7 | |
| CLAIMBackbone=LLaVA-1.52025.06 | 155 | 120 | 125 | 150 | 108.3 | — | — | — | 131.7 | |
| VCDBackbone=Qwen-VL-Chat2025.06 | 150 | — | 165 | 146.7 | — | — | 93.3 | 160 | 143.7 | |
| VCDBackbone=LLaVA-1.52025.06 | 95 | — | 145 | 85 | 130 | 88.3 | — | — | 108.7 | |
| BaselineBackbone=LLaVA-1.52025.06 | 80 | 165 | 135 | 75 | 110 | 80 | — | — | 96 |