List your company

Hallucination Evaluation on MME Existence

190Zh ScoreBaseline
153.6163.05172.5181.95Jun 3, 2025

Evaluation results

MethodLinks
BaselineBackbone=Qwen-VL-Chat2025.06190185135140——106.7195153.3
VCDBackbone=Qwen-VL-Chat2025.06185—155155——78.3195153.7
VCDBackbone=LLaVA-1.52025.06180—130145150128.3——146.7
BaselineBackbone=LLaVA-1.52025.06175190130145155125——146
CLAIMBackbone=LLaVA-1.52025.06175195185175180———182
CLAIMBackbone=Qwen-VL-Chat2025.06155190185128.3——195—170.7