Tracing Training Footprints: A Calibration Approach for Membership Inference Attacks Against Multimodal Large Language Models.

Xiaofan Zheng, Huixuan Zhang, Xiaojun Wan

Browse the full EMNLP paper archive.