AI RESEARCH
Vectors Are Not Neutral: Sensitive-Information Inference from Exported LLM Representations in Summarization
arXiv CS.CL
•
ArXi:2605.26433v1 Announce Type: new Large language model (LLM) summarization systems may pass compact vector representations of private inputs to downstream retrieval, monitoring, audit, or analytic workflows. Even when source documents remain access-restricted, derived vectors may be handled under different access controls and still sensitive-information inference, creating a residual information-disclosure risk. We study this issue in clinical discharge-summary generation as a high-stakes, using electronic health record (EHR)-recorded race as a controlled sensitive-label audit.