List your company

Hallucination evaluation on SPD-Faith Bench

43.8CRSAGE
36.31238.25640.242.144Feb 8, 2026

Evaluation results

MethodLinks
SAGEModel=Qwen2.5-VL-7B2026.0243.833.7
Zoom-RefineModel=Qwen2.5-VL-7B2026.0239.731.5
VCDModel=Qwen2.5-VL-7B2026.0239.431.2
SCModel=Qwen2.5-VL-7B2026.0238.729.8
GreedyModel=Qwen2.5-VL-7B2026.0238.229.5
SRModel=Qwen2.5-VL-7B2026.0237.527.7
APIModel=Qwen2.5-VL-7B2026.0236.628.2