Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts
cs.CR
Submitted: 2026-10-09
Updated: 2026-10-09
Code: https://github.com/ggml-org/llama.cpp
Sources
- Agile-Quant: Activation-Guided Quantization for Faster Inference of LLMs on the Edge
- PhysioLLM: Supporting Personalized Health Insights with Wearables and Large Language Models
- LegalGuardian: A Privacy-Preserving Framework for Secure Integration of Large Language Models in Legal Practice
- Towards Privacy-Preserving Mental Health Support with Large Language Models
- Kraken: Higher-order EM Side-Channel Attacks on DNNs in Near and Far Field
- Real-world Edge Neural Network Implementations Leak Private Interactions Through Physical Side Channel
- The Llama 3 Herd of Models
- Spill The Beans: Exploiting CPU Cache Side-Channels to Leak Tokens from Large Language Models
- Whisper Leak: a side-channel attack on Large Language Models
- Text Embeddings Reveal (Almost) As Much As Text
- Language Model Inversion
- Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
- EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks
- Rethinking the Inception Architecture for Computer Vision
- Adam: A Method for Stochastic Optimization
- WildChat: 1M ChatGPT Interaction Logs in the Wild
- MS MARCO: A Human Generated MAchine Reading COmprehension Dataset
- Super-NaturalInstructions: Generalization via Declarative Instructions on 1600+ NLP Tasks
- ChatDoctor: A Medical Chat Model Fine-Tuned on a Large Language Model Meta-AI (LLaMA) Using Medical Domain Knowledge
- SQuAD: 100,000+ Questions for Machine Comprehension of Text
Related papers
- SoK: AI-Augmented Binary Reversing
- Relaxed Sender Anonymity for CBDC Interbank Settlement: A Zero-Knowledge Approach on Permissioned EVM
- Calibration-Family Overfit: Why Trusted Sabotage Monitors Don't Transfer Across Lineages
- Efficient Fuzzy PSI under One-Sided Assumptions
- Sealing the Audit-Runtime Gap for LLM Skills
- Token Composition: A Graph Based on EVM Logs