Unsafe by Reciprocity: How Generation-Understanding Coupling Undermines Safety in Unified Multimodal Models
arXiv:2603.27332v1 Announce Type: new
Abstract: Recent advances in Large Language Models (LLMs) and Text-to-Image (T2I) models have led to the emergence of Unified Multimodal Models (UMMs), where multimodal understanding and image generation are tight…