List your company
ResearchFormerly WizWand

Self-evaluation on CVBench (AUROC)

0.747AUROCVAUQ
0.47140.542950.61450.68605Feb 24, 2026
Updated 8mo ago

Evaluation results

MethodLinks
VAUQBackbone=InternVL3.5-8B2026.020.747
Semantic EntropyBackbone=InternVL3.5-8B2026.020.737
VAUQModel=LLaVA-1.5-7B2026.020.732
VL-UncertaintyBackbone=InternVL3.5-8B2026.020.72
VL-UncertaintyModel=LLaVA-1.5-7B2026.020.711
Semantic EntropyModel=LLaVA-1.5-7B2026.020.702
VAUQBackbone=Qwen-2.5-VL-7B2026.020.698
VL-UncertaintyBackbone=Qwen-2.5-VL-7B2026.020.697
VAUQModel=LLaVA-1.5-13B2026.020.683
VL-UncertaintyModel=LLaVA-1.5-13B2026.020.679
PerplexityBackbone=InternVL3.5-8B2026.020.664
Semantic EntropyModel=LLaVA-1.5-13B2026.020.662
EigenScoreModel=LLaVA-1.5-7B2026.020.658
PerplexityBackbone=Qwen-2.5-VL-7B2026.020.648
PerplexityModel=LLaVA-1.5-13B2026.020.646
SVARBackbone=Qwen-2.5-VL-7B2026.020.616
EigenScoreBackbone=InternVL3.5-8B2026.020.61
PerplexityModel=LLaVA-1.5-7B2026.020.603
VerbalizedBackbone=InternVL3.5-8B2026.020.601
Contextual LensModel=LLaVA-1.5-13B2026.020.586
EigenScoreModel=LLaVA-1.5-13B2026.020.581
VerbalizedModel=LLaVA-1.5-7B2026.020.576
VerbalizedBackbone=Qwen-2.5-VL-7B2026.020.563
VerbalizedModel=LLaVA-1.5-13B2026.020.554
Contextual LensBackbone=Qwen-2.5-VL-7B2026.020.544
Chain-of-EmbeddingsBackbone=InternVL3.5-8B2026.020.536
Chain-of-EmbeddingsModel=LLaVA-1.5-13B2026.020.533
Contextual LensModel=LLaVA-1.5-7B2026.020.511
EigenScoreBackbone=Qwen-2.5-VL-7B2026.020.509
Semantic EntropyBackbone=Qwen-2.5-VL-7B2026.020.509
Chain-of-EmbeddingsModel=LLaVA-1.5-7B2026.020.506
SVARBackbone=InternVL3.5-8B2026.020.505
Chain-of-EmbeddingsBackbone=Qwen-2.5-VL-7B2026.020.497
SVARModel=LLaVA-1.5-13B2026.020.492
SVARModel=LLaVA-1.5-7B2026.020.487
Contextual LensBackbone=InternVL3.5-8B2026.020.482