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

Paper recorded by Signals 4 on 2026-09-24 in cs.CL. Abstract reproduced from arXiv; link to the original below.

Published 2026-09-24 on arXiv · recorded by Signals 4 on 2026-09-25

Category: cs.CL · 自然语言处理 · first seen 2026-09-25

Abstract

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-

Read on arXiv →