List your company

Object Hallucination Detection on MSCOCO

89.62AUROCInsLen
56.02864.74973.4782.191May 12, 2026

Evaluation results

MethodLinks
InsLenModel=LLaVA-1.5-13B2026.0589.62—————96.27
InsLenModel=InstructBLIP-13B2026.0589.04—————95.46
InsLenModel=LLaVA-1.5-7B2026.0586.93—————96.38
InsLenModel=InstructBLIP-7B2026.0585.5—————94.74
GLSIMModel=LLaVA-1.5-13B2026.0585.36—————95.37
GLSIMModel=InstructBLIP-13B2026.0584.77—————95.01
GLSIMModel=LLaVA-1.5-7B2026.0583.24—————93.81
InsLenModel=LLaVA-OneVision1.5-8B2026.0582.41—————96.99
GLSIMModel=InstructBLIP-7B2026.0582.38—————93.57
Internal Conf.Model=InstructBLIP-7B2026.0581.62—————93.27
InsLenModel=Qwen3-VL-8B2026.0581.02—————96.7
NLLModel=InstructBLIP-7B2026.0578.41—————90.57
SVARModel=InstructBLIP-7B2026.0577.59—————92.63
SVARModel=Qwen3-VL-8B2026.0575.36—————93.44
InsLenModel=mPLUG-Owl3-8B2026.0575.11—————96.47
SVARModel=LLaVA-1.5-7B2026.0573.97—————91.33
Contextual LensModel=InstructBLIP-7B2026.0573.91—————93.04
Internal Conf.Model=LLaVA-1.5-13B2026.0573.66—————91.67
Internal Conf.Model=LLaVA-1.5-7B2026.0573.45—————91.21
GLSIMModel=Qwen3-VL-8B2026.0573.32—————93.41
SVARModel=LLaVA-OneVision1.5-8B2026.0573.08—————95.13
Internal Conf.Model=InstructBLIP-13B2026.0572.34—————89.9
Internal Conf.Model=LLaVA-OneVision1.5-8B2026.0571.89—————93.07
EASYModel=Qwen3-VL-8B2026.0571.02—————92.89
Contextual LensModel=LLaVA-1.5-7B2026.0570.83—————87.88
EASYModel=LLaVA-OneVision1.5-8B2026.0570.43—————93.37
SVARModel=mPLUG-Owl3-8B2026.0570.36—————95.66
EASYModel=LLaVA-1.5-7B2026.0568.05—————85.29
GLSIMModel=LLaVA-OneVision1.5-8B2026.0568.03—————93.24
EntropyModel=InstructBLIP-7B2026.0567.78—————80.73
Internal Conf.Model=Qwen3-VL-8B2026.0567.68—————92.03
GLSIMModel=mPLUG-Owl3-8B2026.0567.12—————94.31
EASYModel=mPLUG-Owl3-8B2026.0566.3—————92.43
EntropyModel=Qwen3-VL-8B2026.0566.23—————91.89
NLLModel=mPLUG-Owl3-8B2026.0566.14—————93.27
EASYModel=InstructBLIP-7B2026.0565.81—————79
EntropyModel=LLaVA-OneVision1.5-8B2026.0565.56—————86.79
NLLModel=LLaVA-1.5-7B2026.0565.19—————84.15
Internal Conf.Model=mPLUG-Owl3-8B2026.0565.07—————94.4
NLLModel=LLaVA-OneVision1.5-8B2026.0564.33—————87.2
NLLModel=Qwen3-VL-8B2026.0564.12—————91.49
Contextual LensModel=Qwen3-VL-8B2026.0563.35—————88.43
EntropyModel=LLaVA-1.5-7B2026.0563.29—————82.21
Contextual LensModel=LLaVA-OneVision1.5-8B2026.0561.02—————92.61
Contextual LensModel=mPLUG-Owl3-8B2026.0557.55—————91.61
EntropyModel=mPLUG-Owl3-8B2026.0557.32—————88.83
EFUFBase Model=LLaVA-v1.5-7B, Setting=Origin2026.01—41.818.3247.785.9—
EFUFBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 602026.01—45.819.72610.485.3—
EFUFBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 1002026.01—46.519.6281184.3—
EFUFBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 1402026.01—48.920.92911.884—
EFUFBase Model=LLaVA-v1.5-7B, Setting=w/ LoRA FT2026.01—48.220.82710.886.1—
EFUFBase Model=LLaVA-v1.5-7B, Setting=w/ Adversarial2026.01—50.523.9281168.5—
EFUFBase Model=mPLUG-Owl-7B, Setting=Origin2026.01—43.624.24617.783.8—
EFUFBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 602026.01—50.526.35018.584.8—
EFUFBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 1002026.01—55.228.65218.982.1—
EFUFBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 1402026.01—59.130.4532081.4—
EFUFBase Model=mPLUG-Owl-7B, Setting=w/ LoRA FT2026.01—58.629.75220.382—
EFUFBase Model=mPLUG-Owl-7B, Setting=w/ Adversarial2026.01—5026.94818.574—
SAREBase Model=LLaVA-v1.5-7B, Setting=Origin2026.01—34.414.7216.586.7—
SAREBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 602026.01—34.614.9216.786.3—
SAREBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 1002026.01—35.415.3227.986.2—
SAREBase Model=LLaVA-v1.5-7B, Setting=w/ Relearn 1402026.01—34.614.8217.486.1—
SAREBase Model=LLaVA-v1.5-7B, Setting=w/ LoRA FT2026.01—37.917.4249.586.4—
SAREBase Model=LLaVA-v1.5-7B, Setting=w/ Adversarial2026.01—39.917.8249.867.7—
SAREBase Model=mPLUG-Owl-7B, Setting=Origin2026.01—37.320.83814.282.3—
SAREBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 602026.01—42.321.84016.384.8—
SAREBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 1002026.01—44.923.84016.884.3—
SAREBase Model=mPLUG-Owl-7B, Setting=w/ Relearn 1402026.01—45.823.94217.784—
SAREBase Model=mPLUG-Owl-7B, Setting=w/ LoRA FT2026.01—47.225.64619.383.8—
SAREBase Model=mPLUG-Owl-7B, Setting=w/ Adversarial2026.01—37.521.23914.771.6—
VanillaBase Model=LLaVA-v1.5-7B, Setting=Baseline2026.01—49.422.14214.785.3—
VanillaBase Model=mPLUG-Owl-7B, Setting=Baseline2026.01—69.633.56024.185.6—