跳到正文
arXiv cs.AI· Prashant Shekhar, Dumindu Samaraweera, Caroline Howard·· 56 分钟前AI 评分50

arXiv 论文推导开放权重 LLM 修改下水印证据的精确界限

Watermark Evidence under Open-Weight LLM Modification: Exact Bounds for Low-Rank Output Watermarks

AI 导读

arXiv 论文 arXiv:2611.21702 研究开放权重 LLM 权重被修改后,OpenStamp 式水印验证器可获得的证据变化,推导出残差约束下期望对数似然比证据的精确有限词表下界,并结合鞅论证给出漏检概率的高置信上界。

正文 · 原文

View PDF HTML (experimental)

Abstract:Open-weight language models can continue generating after their released weights are modified, while a watermark verifier still scores outputs against the original base and marked models. We study how such modifications change the evidence available to an OpenStamp-style verifier. At each shared prefix, we represent the modification as a residual in the next-token logits. We derive an exact finite-vocabulary lower bound on the expected log-likelihood-ratio evidence under a residual-span constraint, along with a sharp span-only deviation bound. We also characterize the first-order evidence shift through a covariance term with an explicit second-order remainder, and determine the residual span needed to reach a target evidence level. For sequence-level detection, we derive an exact conditional exponential-moment bound and combine it with a supermartingale argument and a distribution-free order-statistic tolerance bound. This yields a high-confidence upper bound on miss probability after attack selection on calibration data and evaluation on independent certification continuations. In experiments with OpenStamp on Llama-2-7B and Mistral-7B-v0.3, INT8 quantization and 30% pruning retain test true-positive rates of 85.6-94.8%, while 500-step LoRA fine-tuning lowers them to 28.8% and 11.6%. The covariance prediction tracks the measured mean LoRA evidence drift on both models, with no remainder-bound violations across 160,000 test prefixes. Additionally, we observe zero false positives among 500 base-model continuations per model. Through these theoretical and empirical results, the current research provides a better understanding of watermark robustness to model-editing/tuning attacks.
Subjects: Cryptography and Security (cs.CR); Artificial Intelligence (cs.AI); Computation and Language (cs.CL)
Cite as: arXiv:2611.21702 [cs.CR]
  (or arXiv:2611.21702v1 [cs.CR] for this version)
  https://doi.org/10.48550/arXiv.2611.21702

arXiv-issued DOI via DataCite (pending registration)

Submission history

From: Prashant Shekhar [view email]
[v1] Mon, 16 Nov 2026 15:52:38 UTC (91 KB)

来源:arXiv cs.AI · arxiv.org