A Robust Semantics-based Watermark for Large Language Model against Paraphrasing
Preprint
Abstract
Large language models (LLMs) have show great ability in various natural language tasks. However, there are concerns that LLMs are possible to be used improperly or even illegally. To prevent the malicious usage of LLMs, detecting LLM-generated text becomes crucial in the deployment of LLM applications. Watermarking is an effective strategy to detect the LLM-generated content by encoding a pre-defined secret watermark to facilitate the detection process. However, the majority of existing watermark methods leverage the simple hashes of precedent tokens to partition vocabulary. Such watermark can be easily eliminated by paraphrase and correspondingly the detection effectiveness will be greatly compromised. Thus, to enhance the robustness against paraphrase, we propose a semantics-based watermark framework SemaMark. It leverages the semantics as an alternative to simple hashes of tokens since the paraphrase will likely preserve the semantic meaning of the sentences. Comprehensive experiments are conducted to demonstrate the effectiveness and robustness of SemaMark under different paraphrases.Keywords
All Related Versions
This publication has 10 references indexed in Scilit:
- A Survey on LLM-Generated Text Detection: Necessity, Methods, and Future DirectionsComputational Linguistics, 2025
- A Survey of Text Watermarking in the Era of Large Language ModelsACM Computing Surveys, 2024
- The Science of Detecting LLM-Generated TextCommunications of the ACM, 2024
- On the Generalization of Training-based ChatGPT Detection Methods2024
- GPT-who: An Information Density-based Machine-Generated Text Detector2024
- SeqXGPT: Sentence-Level AI-Generated Text Detection2023
- Creating and detecting fake reviews of online productsJournal of Retailing and Consumer Services, 2022
- All the News That’s Fit to Fabricate: AI-Generated Text as a Tool of Media MisinformationJournal of Experimental Political Science, 2020
- Array programming with NumPyNature, 2020
- A tutorial on kernel density estimation and recent advancesBiostatistics & Epidemiology, 2017