My Answer Is NOT 'Fair': Mitigating Social Bias in Vision-Language Models via Fair and Biased Residuals

Open in new window