paper-with-me

홈 › Papers

Shadow in the Cache: Unveiling and Mitigating Privacy Risks of KV-cache in LLM Inference

2025-08-13 · Zhifan Luo, Shuo Shao, Su Zhang, Lijing Zhou, Yuke Hu, Chenxu Zhao, Zhihao Liu, Zhan Qin arxiv

The Key-Value (KV) cache, which stores intermediate attention computations (Key and Value pairs) to avoid redundant calculations, is a fundamental mechanism for accelerating Large Language Model (LLM) inference. However, this efficiency optimization introduces significant yet underexplored privacy risks. This paper provides the first comprehensive analysis of these vulnerabilities, demonstrating that an attacker can reconstruct sensitive user inputs directly from the KV-cache. We design and implement three distinct attack vectors: a direct Inversion Attack, a more broadly applicable and potent Collision Attack, and a semantic-based Injection Attack. These methods demonstrate the practicality and severity of KV-cache privacy leakage issues. To mitigate this, we propose KV-Cloak, a novel, lightweight, and efficient defense mechanism. KV-Cloak uses a reversible matrix-based obfuscation scheme, combined with operator fusion, to secure the KV-cache. Our extensive experiments show that KV-Cloak effectively thwarts all proposed attacks, reducing reconstruction quality to random noise. Crucially, it achieves this robust security with virtually no degradation in model accuracy and minimal performance overhead, offering a practical solution for trustworthy LLM deployment.

📄 PDF Abstract BibTeX arXiv:2508.09442

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Unveiling Privacy Risks in Multi-modal Large Language Models: Task-specific Vulnerabilities and Mitigation Challenges

2026-06-08 · Tiejin Chen, Pingzhi Li, Kaixiong Zhou, Tianlong Chen 외 arxiv

Privacy risks in text-only Large Language Models (LLMs) are well studied, particularly their tendency to memorize and leak sensitive information. However, Multi-modal Large Language Models (MLLMs), which process both tex…

Privacy Auditing of Multi-domain Graph Pre-trained Model under Membership Inference Attacks

2025-11-22 · Jiayi Luo, Qingyun Sun, Yuecen Wei, Haonan Yuan 외 arxiv

Multi-domain graph pre-training has emerged as a pivotal technique in developing graph foundation models. While it greatly improves the generalization of graph neural networks, its privacy risks under membership inferenc…

Incremental Learning

On the Fairness of Privacy Protection: Measuring and Mitigating the Disparity of Group Privacy Risks for Differentially Private Machine Learning

2025-10-10 · Zhi Yang, Changwu Huang, Ke Tang, Xin Yao arxiv

While significant progress has been made in conventional fairness-aware machine learning (ML) and differentially private ML (DPML), the fairness of privacy protection across groups remains underexplored. Existing studies…

ShadowKV: KV Cache in Shadows for High-Throughput Long-Context LLM Inference

2024-10-28 · Hanshi Sun, Li-Wen Chang, Wenlei Bao, Size Zheng 외

With the widespread deployment of long-context large language models (LLMs), there has been a growing demand for efficient support of high-throughput inference. However, as the key-value (KV) cache expands with the seque…

CPUGPU

Selective KV-Cache Sharing to Mitigate Timing Side-Channels in LLM Inference

2025-08-11 · Kexin Chu, Zecheng Lin, Dawei Xiang, Zixu Shen 외 arxiv

Global KV-cache sharing is an effective optimization for accelerating large language model (LLM) inference, yet it introduces an API-visible timing side channel that lets adversaries infer sensitive user inputs from shar…