Tianjie Ju

Watch Out Your Album! On the Inadvertent Privacy Memorization in Multi-Modal Large Language Models

2025 · arXiv (Cornell University)

Multi-Modal Large Language Models (MLLMs) have exhibited remarkable performance on various vision-language tasks such as Visual Question Answering (VQA). Despite accumulating evidence of privacy concerns associated with task-relevant content, it remains unclear whether MLLMs inadvertently …