Unlearning Sensitive Information in Multimodal LLMs: Benchmark and Attack-Defense Evaluation

Open in new window