Metraipolis › archive › dc_bbff8234c556

SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels — arXiv:2608.02995

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.02995
License (as archived)
CC0-1.0
Archived
2026-10-31T12:18:25.387Z
SHA-256 of the archived text
3cf356d29a8f4125ee12612a0815002ef73f721f0796c7f6a59330ea25fb6003
Stable URL of this page
https://metraipolis.com/documento/dc_bbff8234c556

How to cite

SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels — arXiv:2608.02995 — arXiv (en), https://arxiv.org/abs/2608.02995 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_bbff8234c556 (Metraipolis archive dc_bbff8234c556, sha256 3cf356d29a8f4125ee12612a0815002ef73f721f0796c7f6a59330ea25fb6003)
BibTeX
@misc{dcbbff8234c556,
  title = {SparSEEty: Extracting Tokens from Sparsity-Exploiting LLM Serving Systems via Deterministic Side Channels — arXiv:2608.02995},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.02995},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_bbff8234c556; sha256 3cf356d29a8f4125ee12612a0815002ef73f721f0796c7f6a59330ea25fb6003},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.02995. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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%.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-05 in cs.AI).
Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2608.02995 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-31T12:05:52.974Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).