List your company

Gender Classification on COCO 100% spurious correlation

77.9Average ScoreBAP
48.46856.10963.7571.391May 11, 2026

Evaluation results

MethodLinks
BAPBackbone=CLIP, Evaluation Protocol=Linear Probing2026.0577.974.7
BAPBackbone=SigLIP2, Evaluation Protocol=Linear Probing2026.0577.875.3
BAPBackbone=SigLIP2, Evaluation Protocol=Zero-Shot2026.0577.571.8
DIALBackbone=CLIP, Evaluation Protocol=Zero-Shot2026.0575.153.8
Data-Matched ControlBackbone=SigLIP2, Evaluation Protocol=Trained Evaluation2026.0574.863.5
BAPBackbone=CLIP, Evaluation Protocol=Zero-Shot2026.0574.257.1
RoboShotBackbone=CLIP, Evaluation Protocol=Zero-Shot2026.0573.252.5
Data-Matched ControlBackbone=CLIP, Evaluation Protocol=Trained Evaluation2026.0573.264.6
RoboShotBackbone=SigLIP2, Evaluation Protocol=Zero-Shot2026.0572.967
DIALBackbone=SigLIP2, Evaluation Protocol=Zero-Shot2026.0572.465.2
Native BackboneBackbone=CLIP, Evaluation Protocol=Zero-Shot2026.0571.448.3
Native BackboneBackbone=SigLIP2, Evaluation Protocol=Zero-Shot2026.0571.364.2
WiSE-FTBackbone=CLIP, Evaluation Protocol=Trained Evaluation2026.0565.532.7
ERMBackbone=CLIP, Evaluation Protocol=Linear Probing & Fine-Tuning2026.0563.14.9
ERMBackbone=SigLIP2, Evaluation Protocol=Linear Probing & Fine-Tuning2026.0555.36.2
WiSE-FTBackbone=SigLIP2, Evaluation Protocol=Trained Evaluation2026.0554.936.1
PruSCBackbone=SigLIP2, Evaluation Protocol=Trained Evaluation2026.0554.613.7
PruSCBackbone=CLIP, Evaluation Protocol=Trained Evaluation2026.055311.5
Native BackboneBackbone=SigLIP2, Evaluation Protocol=Linear Probing2026.0551.49.3
Native BackboneBackbone=CLIP, Evaluation Protocol=Linear Probing2026.0549.65.3