A Statistical Framework of Watermarks for Large Language Models: Pivot, Detection Efficiency and Optimal Rules

arXiv:2404.01245v1 Announce Type: cross
Abstract: Since ChatGPT was introduced in November 2022, embedding (nearly) unnoticeable statistical signals into text generated by large language models (LLMs), also known as watermarking, has been used as a principled approach to provable detection of LLM-generated text from its human-written counterpart. In this paper, we introduce a general and flexible framework for reasoning about the statistical efficiency of watermarks and designing powerful detection rules. Inspired by the hypothesis testing formulation of watermark detection, our framework …

arxiv chatgpt cs.cl cs.cr cs.lg detection efficiency framework generated human language language models large llm llms math.st november pivot rules signals stat.ml stat.th text watermarking watermarks written