Alignment in Multimodal Large Language Models: A Survey
Published in Preprints.org, 2026
This survey distinguishes cross-modal evidence alignment from alignment with human goals and constraints. It examines architecture, objectives, data, training, evaluation, and applications, and outlines directions for evidence-consistent alignment and reliable self-correction.
First posted on October 14, 2026; author information follows the updated version posted on October 20, 2026.
Recommended citation: Yuyao Yang, Jiaxing Zhang, Ling Yue, Linsey Pang, Zhiwei Liu, Zixian Ma, Zheng Zhao, Jialiang Yin, Mingdai Yang, Huayi Yang, Lucas Ni, Yi-Fan Zhang, Connor Lee, Jiayuan Ding, Ranjay Krishna, Bin Shi, Zenglin Xu, Yan Liu, Qingsong Wen, Philip S. Yu. 2026. Alignment in Multimodal Large Language Models: A Survey. Preprints.org, 202610.1245 (version 2).
Recommended citation: Yuyao Yang, Jiaxing Zhang, Ling Yue, Linsey Pang, Zhiwei Liu, Zixian Ma, Zheng Zhao, Jialiang Yin, Mingdai Yang, Huayi Yang, Lucas Ni, Yi-Fan Zhang, Connor Lee, Jiayuan Ding, Ranjay Krishna, Bin Shi, Zenglin Xu, Yan Liu, Qingsong Wen, Philip S. Yu. 2026. Alignment in Multimodal Large Language Models: A Survey. Preprints.org, 202610.1245 (version 2). https://www.preprints.org/manuscript/202610.1245