List your company

Multimodal Safety Evaluation on GOAT (test)

56.9Misogyny AccuracyOSGA
49.151.12553.1555.175Jan 30, 2026

Evaluation results

MethodLinks
OSGABase Model=LLaVA-v1.52026.0156.936.7147.5157.6150.3838.368.0247.1255.744.94
BaselineBase Model=LLaVA-v1.52026.0149.432.7641.5930.4850.1133.584434.5246.2832.84