Newest papers

A Unified Attack Framework against Statistical Generative Watermarks for Large Language Models

Romina Omidi, Nima Naderloui, Yuan Hong, Binghui Wang · 2026 · arXiv

DOI: 10.48550/arXiv.2610.27653 · open-access PDF

Chinese

水标记已成为大型语言模型(LLM)生成的文本识别的关键工具,目前的方案将统计信号嵌入生成过程中,以便允许后续检测。在这项工作中,我们确定了广泛的一类流行的LLM水标记方法的共同弱点,包括KGW(及其变体),GumbelSoft,DipMark,ITS和SynthID。我们展示了这些方案在检测分数和它们在生成过程中引入的代币概率变化之间存在根本联系。

Machine translation (Meta M2M100); the English original follows.

Abstract

Watermarking has emerged as a key tool for identifying text generated by large language models (LLMs), with current schemes embedding statistical signals into the generation process to enable later detection. In this work, we identify a shared weakness across a broad class of popular LLM watermarking methods, including KGW (and its variants), GumbelSoft, DipMark, ITS, and SynthID. We show that these schemes exhibit a fundamental link between their detection scores and the token probability shifts they introduce during generation. This link, which we call score-ratio alignment, allows an attack

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange