Preprint · 2026 · Open access
Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts
arXiv (Cornell University) · Published 9 October 2026 · DOI 10.48550/arxiv.2610.14215
In brief
“Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts” is a 2026 preprint by Najeh Alhalawani and 3 co-authors, published in arXiv (Cornell University). OpenAlex counts 0 citations as of 19 November 2026. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).
- Citations
- 0
- References
- 49
- Authors
- 4
Abstract
The deployment of large language models (LLMs) on edge devices is largely motivated by privacy concerns over sharing sensitive user data with cloud servers. In this paper, we demonstrate that keeping inference local does not eliminate privacy risks: an adversary with physical access to the device may exploit hardware leakage to recover sensitive information without compromising or modifying the software system. In particular, we present an electromagnetic (EM) side-channel attack targeting LLM inference on edge devices. Targeting LLaMA-3.2-1B-Instruct running on a Raspberry Pi 5, we demonstrate that EM traces collected during inference can leak both generated tokens and user prompts. We propose three attacks under different threat assumptions. First, a profiling attack that identifies generated tokens among 50 medical and privacy-related keywords with 97.2% accuracy on profiled words. Second, a reconstruction attack recovers the model's hidden representations and uses them to predict tokens from the full 128k-token vocabulary, achieving 94.3% accuracy on profiled classes. For unseen words, the correct token is ranked within the first 100 positions of the 128k-token vocabulary in 67% of cases, with a median rank of 54. Finally, we present a hidden-state inversion attack that uses the recovered hidden state to reconstruct user prompts semantically. Our proposed attacks require only passive EM side-channel measurements and a limited number of traces during model training. These results demonstrate that local execution alone may not guarantee the confidentiality of generated tokens and user inputs.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Alhalawani, N., Acar, G., Liu, Z., & Batina, L. (2026). Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.14215
MLA 9
Alhalawani, Najeh, et al. “Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2610.14215.
Chicago
Alhalawani, Najeh, Güneş Acar, Zhuoran Liu, and Lejla Batina. 2026. “Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.14215.
Harvard
Alhalawani, N. et al. (2026) ‘Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts’, arXiv (Cornell University). doi:10.48550/arxiv.2610.14215.
IEEE
N. Alhalawani, G. Acar, Z. Liu, and L. Batina, “Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2610.14215.
BibTeX
@misc{alhalawani2026physical,
title = {{Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts}},
author = {Alhalawani, Najeh and Acar, Güneş and Liu, Zhuoran and Batina, Lejla},
howpublished = {arXiv (Cornell University)},
year = {2026},
doi = {10.48550/arxiv.2610.14215},
url = {https://doi.org/10.48550/arxiv.2610.14215}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Hardware and Architecture › Security and Verification in Computing
- electromagnetic side-channel
- prompt reconstruction
- llamas
- Raspberry Pi
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Alhalawani, N., Acar, G., Liu, Z., & Batina, L. (2026). Physical Side-Channel Attacks on Edge LLMs: Extracting Output Tokens and User Prompts. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2610.14215
How many times has this paper been cited?
0 times, according to OpenAlex on 19 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access at arXiv (Cornell University): sign in and open the PDF from this page.
Who wrote this paper?
Najeh Alhalawani, Güneş Acar, Zhuoran Liu, Lejla Batina.
What is the DOI of this paper?
10.48550/arxiv.2610.14215, permanently at https://doi.org/10.48550/arxiv.2610.14215. Sign in and open it from this page.
Bibliographic data from OpenAlex (CC0), last checked 19 November 2026. ResearchGuru does not host this paper; once you are signed in, the links open it at the publisher or the repository that does.