Identifying and Mitigating Position Bias of Multi-image Vision-Language Models.

Xinyu Tian, Shu Zou, Zhaoyuan Yang, Jing Zhang

Browse the full CVPR paper archive.