Jiang, W. et al. 2026. PurMM: Attention-Guided Test-Time Backdoor Purification in Multimodal Large Language Models. Proceedings of the AAAI Conference on Artificial Intelligence. 40, 42 (Mar. 2026), 35562–35570. DOI:https://doi.org/10.1609/aaai.v40i42.40867.