Source · Tier B · Preprint
Hollow-LLM Attack: Computationally Trivial Weights in Zero-Knowledge Verification of LLM Inference
C. Gong, B. Liu, M. Li. 2026. arXiv.
| Link | https://arxiv.org/abs/2607.28884 |
|---|---|
| arXiv | 2607.28884 |
| Version | v1 (30 July 2026), read via the arXiv HTML and PDF renderings; the abstract page did not render for the fetch tool. Author order is as printed on the paper's title page; a mirror of the arXiv listing (pith.science) lists B. Liu first. A co-author's web page labels the paper "IEEE S&P 2026", but no proceedings record was found and a co-author's CV lists it as under submission, so it is recorded as a preprint. |
| Accessed | 2026-09-24 |
| Note | Independent analysis (University of Southern California) of zero-knowledge proofs of LLM inference. Shows that valid proofs do not bind the computation spent, using "ghost weights"; experiments use the proof procedure of zkGPT on a 6-layer, 512-dimensional transformer. |