List your company

Multi-view Hallucination Evaluation on MVH-Bench Cross-Instance 1.0 (test)

67.5Binary AccuracyRSCD
57.994460.462262.9365.3978Mar 25, 2026

Evaluation results

MethodLinks
RSCDBackbone=Qwen2.5-VL-7B2026.0367.545.1917.6973.6653.7257.73
VCDBackbone=Qwen2.5-VL-7B2026.0367.4345.7919.4471.7150.28254.65
AvisCBackbone=Qwen2.5-VL-7B2026.0367.0645.092073.853.43259.38
ICDBackbone=Qwen2.5-VL-7B2026.0366.5343.818.772.7352.59254.35
Base modelBackbone=Qwen2.5-VL-7B2026.0364.141.316.2966.5743.52231.78
RSCDBackbone=LLaVA-OneVision-7B2026.0362.6638.114.4567.6444.91227.76
VCDBackbone=LLaVA-OneVision-7B2026.0360.8632.7310.9368.1544.26216.92
ICDBackbone=LLaVA-OneVision-7B2026.0360.4633.5212.2267.2242.78216.2
AvisCBackbone=LLaVA-OneVision-7B2026.0360.443010.1967.7343.15211.51
Base modelBackbone=LLaVA-OneVision-7B2026.0358.3632.3610.9362.7836.29200.72