The Coherence Trap: When MLLM-Crafted Narratives Exploit Manipulated Visual Contexts
Published in CVPR 2026, 2026
Status: Accepted to CVPR 2026 (main track).
Recommended citation: Zhang, Y., Wang, Y., Wu, Y., Wu, L., Zhu, L., & Zheng, Z. (2026). "The Coherence Trap: When MLLM-Crafted Narratives Exploit Manipulated Visual Contexts." CVPR 2026. Accepted to the main track. Preprint: https://arxiv.org/abs/2505.17476
Download Paper