Red Teaming Multimodal Language Models for Amazon: Evaluating Harm Across Prompt Modalities and Models Read More
