List your company

Object Hallucination Evaluation on POPE GQA (random)

92.23AccuracySIRA
79.146882.543485.9489.3366Jan 3, 2025Mar 26, 2025Jun 17, 2025Sep 8, 2025Nov 29, 2025Feb 20, 2026May 14, 2026

Evaluation results

MethodLinks
SIRABackbone=Qwen2.5-VL2026.0592.2392.09——
CMPCBase Model=LLaVA 1.52025.0189.5388.890.6288.2
SIRABackbone=LLaVA-v1.52026.0589.4789.53——
MaskCDBackbone=Qwen2.5-VL2026.0589.2589.1——
ICTBackbone=LLaVA-v1.52026.0589.2389.34——
ICTBackbone=Qwen2.5-VL2026.0588.9687.89——
OPERABackbone=LLaVA-v1.52026.0588.1388.91——
MaskCDBackbone=LLaVA-v1.52026.0587.8588.25——
DoLaBackbone=Qwen2.5-VL2026.0587.1386.05——
SIDBackbone=LLaVA-v1.52026.0586.9787.17——
VCDBackbone=LLaVA-v1.52026.0586.6586.99——
VCDBase Model=LLaVA 1.52025.0186.6586.9984.8589.24
CMPCBase Model=InstructBLIP2025.0186.0385.7287.6783.87
OPERABackbone=Qwen2.5-VL2026.0586.0285.29——
ICDBase Model=LLaVA 1.52025.0185.7384.7990.7979.53
VCDBackbone=Qwen2.5-VL2026.0585.5985.33——
BaseBackbone=Qwen2.5-VL2026.0585.183.87——
ICDBase Model=InstructBLIP2025.0185.185.2984.2186.4
SIDBackbone=Qwen2.5-VL2026.0584.8684.25——
BaseBackbone=LLaVA-v1.52026.0583.7382.95——
BaselineBase Model=LLaVA 1.52025.0183.7382.9587.1679.12
DoLaBackbone=LLaVA-v1.52026.0583.785.29——
VCDBase Model=InstructBLIP2025.0183.6984.1681.8486.61
BaselineBase Model=InstructBLIP2025.0179.6580.5677.1484.29