Watermark under Fire: A Robustness Evaluation of LLM Watermarking

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Liang, Jiacheng, Wang, Zian, Hong, Lauren, Ji, Shouling, Wang, Ting
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909814943121408
author Liang, Jiacheng
Wang, Zian
Hong, Lauren
Ji, Shouling
Wang, Ting
author_facet Liang, Jiacheng
Wang, Zian
Hong, Lauren
Ji, Shouling
Wang, Ting
contents Various watermarking methods (``watermarkers'') have been proposed to identify LLM-generated texts; yet, due to the lack of unified evaluation platforms, many critical questions remain under-explored: i) What are the strengths/limitations of various watermarkers, especially their attack robustness? ii) How do various design choices impact their robustness? iii) How to optimally operate watermarkers in adversarial environments? To fill this gap, we systematize existing LLM watermarkers and watermark removal attacks, mapping out their design spaces. We then develop WaterPark, a unified platform that integrates 10 state-of-the-art watermarkers and 12 representative attacks. More importantly, by leveraging WaterPark, we conduct a comprehensive assessment of existing watermarkers, unveiling the impact of various design choices on their attack robustness. We further explore the best practices to operate watermarkers in adversarial environments. We believe our study sheds light on current LLM watermarking techniques while WaterPark serves as a valuable testbed to facilitate future research.
format Preprint
id arxiv_https___arxiv_org_abs_2411_13425
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Watermark under Fire: A Robustness Evaluation of LLM Watermarking
Liang, Jiacheng
Wang, Zian
Hong, Lauren
Ji, Shouling
Wang, Ting
Cryptography and Security
Computation and Language
Machine Learning
Various watermarking methods (``watermarkers'') have been proposed to identify LLM-generated texts; yet, due to the lack of unified evaluation platforms, many critical questions remain under-explored: i) What are the strengths/limitations of various watermarkers, especially their attack robustness? ii) How do various design choices impact their robustness? iii) How to optimally operate watermarkers in adversarial environments? To fill this gap, we systematize existing LLM watermarkers and watermark removal attacks, mapping out their design spaces. We then develop WaterPark, a unified platform that integrates 10 state-of-the-art watermarkers and 12 representative attacks. More importantly, by leveraging WaterPark, we conduct a comprehensive assessment of existing watermarkers, unveiling the impact of various design choices on their attack robustness. We further explore the best practices to operate watermarkers in adversarial environments. We believe our study sheds light on current LLM watermarking techniques while WaterPark serves as a valuable testbed to facilitate future research.
title Watermark under Fire: A Robustness Evaluation of LLM Watermarking
topic Cryptography and Security
Computation and Language
Machine Learning
url https://arxiv.org/abs/2411.13425