Fetching the paper…

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