LLMs as Linguistic Chameleons: Decoupling Semantics and Structure for Privacy-Preserving Communication

2026-09-19

AI Summary

本文提出CROSS-MAP框架，通过将隐私输入映射到不同语义域并在推理后恢复输出，以解决大型语言模型在保持任务实用性的同时保护隐私的问题。

Abstract

As Large Language Model (LLM) APIs become increasingly integrated into privacy-sensitive workflows, ensuring inference-time privacy without compromising task utility remains a major challenge. Existing approaches preserve most of the original semantic content to maintain downstream performance, but this also leaves exploitable cues for reconstructing the original text. This work investigates semantic decoupling, which replaces original semantics with alternative content while preserving the structure needed for LLM reasoning. Based on this idea, we propose CROSS-MAP, a bidirectional framework that maps private inputs into a different semantic domain before inference and recovers the corresponding outputs afterward. Local models are trained with multi-objective optimization to maximize semantic divergence in the mapping stage while minimizing semantic inconsistency in the recovery stage. Experiments show that CROSS-MAP reduces reconstruction success across multiple attack settings while outperforming existing baselines in utility.

Problem

• privacy-preserving communication
• large language models
• semantic decoupling
• inference-time privacy

Innovation

• semantic decoupling
• CROSS-MAP
• multi-objective optimization
• privacy-preserving