Paper · arXiv 2308.00113
Three Bricks to Consolidate Watermarks for Large Language Models
Pierre Fernandez, Antoine Chaffin, Karim Tit, Vivien Chappelier, Teddy Furon
This research develops robust watermarking techniques for large language models, introducing statistical tests, comparing watermarks through classical benchmarks, and enhancing detection schemes.
Abstract
The task of discerning between generated and natural texts is increasingly challenging. In this context, watermarking emerges as a promising technique for ascribing generated text to a specific model. It alters the sampling generation process so as to leave an invisible trace in the generated output, facilitating later detection. This research consolidates watermarks for large language models based on three theoretical and empirical considerations. First, we introduce new statistical tests that offer robust theoretical guarantees which remain valid even at low false-positive rates (less than 10^{-6}). Second, we compare the effectiveness of watermarks using classical benchmarks in the field of natural language processing, gaining insights into their real-world applicability. Third, we develop advanced detection schemes for scenarios where access to the LLM is available, as well as multi-bit watermarking.