BadToken: Token-level Backdoor Attacks to Multi-modal Large Language Models.

Zenghui Yuan, Jiawen Shi, Pan Zhou, Neil Zhenqiang Gong, Lichao Sun

Browse the full CVPR paper archive.