Copertina

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

Fuente: arXiv

Salvato in:

Dettagli Bibliografici
Autori principali:	Luo, Zhifan, Shao, Shuo, Zhang, Su, Zhou, Lijing, Hu, Yuke, Zhao, Chenxu, Liu, Zhihao, Qin, Zhan
Natura:	Preprint
Pubblicazione:	2025
Soggetti:	Cryptography and Security Artificial Intelligence Computation and Language
Accesso online:	Acceder al recurso
Tags:	Aggiungi Tag Nessun Tag, puoi essere il primo ad aggiungerne!!

Documenti analoghi

CachePrune: Privacy-Aware and Fine-Grained KV Cache Sharing for Efficient LLM Inference
di: Wu, Guanlong, et al.
Pubblicazione: (2026)

Selective KV-Cache Sharing to Mitigate Timing Side-Channels in LLM Inference
di: Chu, Kexin, et al.
Pubblicazione: (2025)

MPCache: MPC-Friendly KV Cache Eviction for Efficient Private LLM Inference
di: Zeng, Wenxuan, et al.
Pubblicazione: (2025)

Membership Inference Attacks Against Vision-Language Models
di: Hu, Yuke, et al.
Pubblicazione: (2025)

CacheSquash: Making caches speculation-aware
di: ElAtali, Hossam, et al.
Pubblicazione: (2024)

Cachemir: Fully Homomorphic Encrypted Inference of Generative Large Language Model with KV Cache
di: Yu, Ye, et al.
Pubblicazione: (2026)

AttriGuard: Defeating Indirect Prompt Injection in LLM Agents via Causal Attribution of Tool Invocations
di: He, Yu, et al.
Pubblicazione: (2026)

Unveiling Privacy Risks in LLM Agent Memory
di: Wang, Bo, et al.
Pubblicazione: (2025)

SWAT: A System-Wide Approach to Tunable Leakage Mitigation in Encrypted Data Stores
di: Zheng, Leqian, et al.
Pubblicazione: (2023)

Profiling for Pennies: Unveiling the Privacy Iceberg of LLM Agents
di: Chen, Jiahao, et al.
Pubblicazione: (2026)

Mitigating Privacy Risk in Membership Inference by Convex-Concave Loss
di: Liu, Zhenlong, et al.
Pubblicazione: (2024)

Collaborative Shadows: Distributed Backdoor Attacks in LLM-Based Multi-Agent Systems
di: Zhu, Pengyu, et al.
Pubblicazione: (2025)

Bicoptor 2.0: Addressing Challenges in Probabilistic Truncation for Enhanced Privacy-Preserving Machine Learning
di: Zhou, Lijing, et al.
Pubblicazione: (2023)

ERASER: Machine Unlearning in MLaaS via an Inference Serving-Aware Approach
di: Hu, Yuke, et al.
Pubblicazione: (2023)

Quantifying and Defending against Privacy Threats on Federated Knowledge Graph Embedding
di: Hu, Yuke, et al.
Pubblicazione: (2023)

CryptoGen: Secure Transformer Generation with Encrypted KV-Cache Reuse
di: Zhang, Hedong, et al.
Pubblicazione: (2026)

Shadow-Free Membership Inference Attacks: Recommender Systems Are More Vulnerable Than You Thought
di: Chi, Xiaoxiao, et al.
Pubblicazione: (2024)

Privacy in Action: Towards Realistic Privacy Mitigation and Evaluation for LLM-Powered Agents
di: Wang, Shouju, et al.
Pubblicazione: (2025)

Accelerating Suffix Jailbreak attacks with Prefix-Shared KV-cache
di: Wang, Xinhai, et al.
Pubblicazione: (2026)

CachePrune: Teaching LLMs What Not to Follow via KV-Cache Editing
di: Wang, Rui, et al.
Pubblicazione: (2025)

Bicoptor: Two-round Secure Three-party Non-linear Computation without Preprocessing for Privacy-preserving Machine Learning
di: Zhou, Lijing, et al.
Pubblicazione: (2022)

Visualizing the Shadows: Unveiling Data Poisoning Behaviors in Federated Learning
di: Zhang, Xueqing, et al.
Pubblicazione: (2024)

Bots can Snoop: Uncovering and Mitigating Privacy Risks of Bots in Group Chats
di: Chou, Kai-Hsiang, et al.
Pubblicazione: (2024)

CacheTrap: Unveiling a Stealthier Gray-Box Trojan against LLMs
di: Nahian, Mohaiminul Al, et al.
Pubblicazione: (2025)

Whose Narrative is it Anyway? A KV Cache Manipulation Attack
di: Ganesh, Mukkesh, et al.
Pubblicazione: (2025)

Bit-Flip Vulnerability of Shared KV-Cache Blocks in LLM Serving Systems
di: Yamamoto, Yuji, et al.
Pubblicazione: (2026)

SmartGuard: Leveraging Large Language Models for Network Attack Detection through Audit Log Analysis and Summarization
di: Zhang, Hao, et al.
Pubblicazione: (2025)

Toward Efficient Inference Attacks: Shadow Model Sharing via Mixture-of-Experts
di: Bai, Li, et al.
Pubblicazione: (2025)

PRISM: Privacy-Aware Routing for Adaptive Cloud-Edge LLM Inference via Semantic Sketch Collaboration
di: Zhan, Junfei, et al.
Pubblicazione: (2025)

Exposing LLM User Privacy via Traffic Fingerprint Analysis: A Study of Privacy Risks in LLM Agent Interactions
di: Zhang, Yixiang, et al.
Pubblicazione: (2025)

Privacy Leaks by Adversaries: Adversarial Iterations for Membership Inference Attack
di: Xue, Jing, et al.
Pubblicazione: (2025)

MIRAGE: Misleading Retrieval-Augmented Generation via Black-box and Query-agnostic Poisoning Attacks
di: Chen, Tailun, et al.
Pubblicazione: (2025)

KV-Auditor: Auditing Local Differential Privacy for Correlated Key-Value Estimation
di: Xu, Jingnan, et al.
Pubblicazione: (2025)

RerouteGuard: Understanding and Mitigating Adversarial Risks for LLM Routing
di: Zhang, Wenhui, et al.
Pubblicazione: (2026)

Chasing Shadows: Pitfalls in LLM Security Research
di: Evertz, Jonathan, et al.
Pubblicazione: (2025)

A First Look At Efficient And Secure On-Device LLM Inference Against KV Leakage
di: Yang, Huan, et al.
Pubblicazione: (2024)

Unveiling Usability Challenges in Web Privacy Controls
di: Masood, Rahat, et al.
Pubblicazione: (2025)

MAJIC: Markovian Adaptive Jailbreaking via Iterative Composition of Diverse Innovative Strategies
di: Qi, Weiwei, et al.
Pubblicazione: (2025)

Differential Privacy in Federated Learning: Mitigating Inference Attacks with Randomized Response
di: Ozturk, Ozer, et al.
Pubblicazione: (2025)

Beyond Jailbreak: Unveiling Risks in LLM Applications Arising from Blurred Capability Boundaries
di: Zhang, Yunyi, et al.
Pubblicazione: (2025)