Breaking the "Provable Security": Detecting Finite-Precision Artifacts in LLM-based Steganography via Low-Probability Vanishing

Wenzhao Cao, Yaofei Wang, Donghui Hu


Abstract
Recent advances in Large Language Models have fostered a new class of generative linguistic steganography, claim “provably secure” by theoretically aligning the steganographic distribution with the language model’s natural distribution. We challenge this premise by exposing Low-Probability Vanishing (LPV), an inevitable vulnerability arising from finite-precision arithmetic. To exploit this, we propose RRNs-HT, a novel steganalysis framework based on Representative Random Numbers and Hypothesis Testing, which transforms the detection task from semantic classification to a statistical audit of the sampling mechanism. Crucially, unlike previous work that contrasts machine text against human text, we validate our method in a rigorous homologous setting to strictly isolate sampling artifacts. Experiments demonstrate that RRNs-HT effectively breaks the security of AC and Meteor with high detection accuracy, whereas state-of-the-art semantic steganalyzers degrade to random guessing. Our findings prove that theoretical security is unattainable in practice without addressing finite-precision leakage.
Anthology ID:
2026.findings-acl.1013
Volume:
Findings of the Association for Computational Linguistics: ACL 2026
Month:
July
Year:
2026
Address:
San Diego, California, United States
Editors:
Maria Liakata, Viviane P. Moreira, Jiajun Zhang, David Jurgens
Venue:
Findings
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
20262–20274
Language:
URL:
https://aclanthology.org/2026.findings-acl.1013/
DOI:
10.18653/v1/2026.findings-acl.1013
Bibkey:
Cite (ACL):
Wenzhao Cao, Yaofei Wang, and Donghui Hu. 2026. Breaking the "Provable Security": Detecting Finite-Precision Artifacts in LLM-based Steganography via Low-Probability Vanishing. In Findings of the Association for Computational Linguistics: ACL 2026, pages 20262–20274, San Diego, California, United States. Association for Computational Linguistics.
Cite (Informal):
Breaking the “Provable Security”: Detecting Finite-Precision Artifacts in LLM-based Steganography via Low-Probability Vanishing (Cao et al., Findings 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.findings-acl.1013.pdf
Checklist:
 2026.findings-acl.1013.checklist.pdf