Paper Overview
- Research areas: cs.AI, cs.CR
- Authors: Zhenhua Liu, Zhanxu Xie, Junjie Yu, Tong Zhu, Lijun Li, Wenliang Chen
- arXiv: 2609.10992
Abstract
The integration of Large Language Models into daily tasks relies on context-rich instructions, inevitably exposing sensitive user information. Current privacy-preserving methods typically employ context-agnostic static rules, causing severe utility degradation. However, the specific mechanisms governing how sanitization impacts downstream performance remain largely underexplored.
To address this, the authors conduct a systematic analysis to deconstruct the privacy-utility trade-off, uncovering three underlying mechanisms:
1. Context-Dependent Utility — establishes *when* to sanitize by revealing that data value shifts from critical constraints to dispensable noise based on user intent. 2. Strategic Adaptation — determines *how* to sanitize by dictating that the choice between removal and replacement depends on the task's reliance on factual integrity versus structural coherence. 3. Combinatorial Interplay — extends the protection scope by demonstrating that attributes form a semantic web of synergistic dependencies or antagonistic redundancies.
Guided by these insights, the paper introduces an intent-driven local protection framework. By distilling a lightweight model, Veilmind-4B, to drive a dynamic extraction-sanitization-restoration pipeline, the approach reaches a low-leakage privacy point while preserving substantially higher response utility than existing privacy-oriented baselines, advancing the privacy-utility trade-off toward the Pareto frontier.
---
*Auto-collected on 2026-09-13*