ETA: Evaluating Then Aligning Safety of Vision Language Models at Inference Time.

Yi Ding, Bolian Li, Ruqi Zhang

Browse the full ICLR paper archive.