Preprint · 2026 · Open access

SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels

조용완, Jinyoung Park, Euihyun Lee, Dokyung Song

arXiv (Cornell University) · Published 4 August 2026 · DOI 10.48550/arxiv.2608.02995

Open PDFView at the publisherCite this paper

In brief

“SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels” is a 2026 preprint by 조용완 and 3 co-authors, published in arXiv (Cornell University). OpenAlex counts 0 citations as of 19 November 2026. It is open access: the full text is free to read at arXiv (Cornell University).

Citations
0
References
70
Authors
4

Abstract

Modern large language models (LLMs) exhibit activation sparsity, wherein only a subset of their neurons is activated for given input tokens. Researchers have leveraged this property to optimize LLM serving systems by omitting weight accesses and computations pertaining to inactive neurons. Unfortunately, however, such optimizations create input-dependent weight accesses, which can be leaked over side channels. We present SparSEEty, a new token extraction attack that exploits input-dependent neuron weight accesses introduced by sparsity-exploiting LLM serving systems. SparSEEty first constructs a neuron-activation oracle using neuron weight access side channels during LLM inference, and then inverts the activation traces to reconstruct the input tokens, forming an end-to-end token extraction attack. We instantiate SparSEEty against an LLM serving system protected inside an Intel TDX confidential virtual machine (CVM), addressing three key challenges: (i) constructing a neuron-activation oracle using a combination of side channels exposed by CVMs, (ii) reducing inference-time overheads of neuron activation monitoring for covertness, and (iii) accurately inverting partial binary activation traces back to tokens. Our evaluation shows that SparSEEty can reconstruct both prompt and response tokens with consistently high BLEU scores (>0.95) across various models and datasets, while incurring monitoring overheads of 3.7% to 7.2%.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

조용완, Park, J., Lee, E., & Song, D. (2026). SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.02995

MLA 9

조용완, et al. “SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels.” arXiv (Cornell University), 2026. https://doi.org/10.48550/arxiv.2608.02995.

Chicago

조용완, Jinyoung Park, Euihyun Lee, and Dokyung Song. 2026. “SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels.” arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.02995.

Harvard

조용완 et al. (2026) ‘SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels’, arXiv (Cornell University). doi:10.48550/arxiv.2608.02995.

IEEE

조용완, J. Park, E. Lee, and D. Song, “SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels,” arXiv (Cornell University), 2026, doi: 10.48550/arxiv.2608.02995.

BibTeX

@misc{2026sparseety,
  title     = {{SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels}},
  author    = {조용완 and Park, Jinyoung and Lee, Euihyun and Song, Dokyung},
  howpublished = {arXiv (Cornell University)},
  year      = {2026},
  doi       = {10.48550/arxiv.2608.02995},
  url       = {https://doi.org/10.48550/arxiv.2608.02995}
}

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

  • large language model serving
  • activation sparsity
  • side channels
  • inference privacy

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

조용완, Park, J., Lee, E., & Song, D. (2026). SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels. arXiv (Cornell University). https://doi.org/10.48550/arxiv.2608.02995

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?

조용완, Jinyoung Park, Euihyun Lee, Dokyung Song.

What is the DOI of this paper?

10.48550/arxiv.2608.02995, permanently at https://doi.org/10.48550/arxiv.2608.02995. 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.