SOURCE-LINKED INTELLIGENCE

PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations

arXiv · AI, language, vision and robotics · article · Sep 24, 2026 · UTC

Large language models increasingly operate as persistent assistants in user-facing, shared-session, and tool-augmented settings. When users disclose sensitive information during an active conversation, that information may remain behaviorally recoverable through later prompts even after the dialogue shifts to unrelated topics. We introduce \textbf{PrivDrift}, a benchmark for auditing whether user-disclosed secrets remain recoverable after conversational topic drift and persuasion-based probing. PrivDrift contains 1{,}000 controlled multi-turn dialogues with seeded secrets, content-dense drift

Read original source ↗ Open in workspace

recordType

paper

region

Global