1 paper
Anqi Li, Jie Zhang, Zhongqi Wang +4
While Large Vision-Language Models (LVLMs) demonstrate remarkable capabilities, they remain highly susceptible to embedded social biases. Existing bias evaluation protocols predomi…