Debiasing Large Vision-Language Models by Ablating Protected Attribute Representations

Re-assign community