all InfoSec news
Proving membership in LLM pretraining data via data watermarks
Feb. 19, 2024, 5:10 a.m. | Johnny Tian-Zheng Wei, Ryan Yixiang Wang, Robin Jia
cs.CR updates on arXiv.org arxiv.org
Abstract: Detecting whether copyright holders' works were used in LLM pretraining is poised to be an important problem. This work proposes using data watermarks to enable principled detection with only black-box model access, provided that the rightholder contributed multiple training documents and watermarked them before public release. By applying a randomly sampled data watermark, detection can be framed as hypothesis testing, which provides guarantees on the false detection rate. We study two watermarks: one that inserts …
access arxiv box contributed copyright cs.cl cs.cr cs.lg data detection documents enable important llm problem public release training watermarks work
More from arxiv.org / cs.CR updates on arXiv.org
Jobs in InfoSec / Cybersecurity
IT Security Manager
@ Timocom GmbH | Erkrath, Germany
Cybersecurity Service Engineer
@ Motorola Solutions | Singapore, Singapore
Sr Cybersecurity Vulnerability Specialist
@ Health Care Service Corporation | Chicago Illinois HQ (300 E. Randolph Street)
Associate, Info Security (SOC) analyst
@ Evolent | Pune
Public Cloud Development Security and Operations (DevSecOps) Manager
@ Danske Bank | Copenhagen K, Denmark
Cybersecurity Risk Analyst IV
@ Computer Task Group, Inc | United States