Robust Multi-bit Text Watermark with LLM-based Paraphrasers

TL;DR

This paper introduces a novel multi-bit text watermarking method using fine-tuned LLM paraphrasers that embeds information imperceptibly and robustly, maintaining semantic integrity and resisting common text modifications.

Contribution

The paper presents a new multi-bit text watermarking technique leveraging dual LLM paraphrasers and a trained decoder, achieving high detection accuracy and robustness.

Findings

Over 99.99% detection AUC with small paraphrasers

Robust against word substitution and sentence paraphrasing

Maintains semantic integrity of original text

Abstract

We propose an imperceptible multi-bit text watermark embedded by paraphrasing with LLMs. We fine-tune a pair of LLM paraphrasers that are designed to behave differently so that their paraphrasing difference reflected in the text semantics can be identified by a trained decoder. To embed our multi-bit watermark, we use two paraphrasers alternatively to encode the pre-defined binary code at the sentence level. Then we use a text classifier as the decoder to decode each bit of the watermark. Through extensive experiments, we show that our watermarks can achieve over 99.99\% detection AUC with small (1.1B) text paraphrasers while keeping the semantic information of the original sentence. More importantly, our pipeline is robust under word substitution and sentence paraphrasing perturbations and generalizes well to out-of-distributional data. We also show the stealthiness of our watermark with LLM-based evaluation. We open-source the code: https://github.com/xiaojunxu/multi-bit-text-watermark.