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
Information Security Engineers
@ D. E. Shaw Research | New York City
Technology Security Analyst
@ Halton Region | Oakville, Ontario, Canada
Senior Cyber Security Analyst
@ Valley Water | San Jose, CA
Sr. Staff Firmware Engineer – Networking & Firewall
@ Axiado | Bengaluru, India
Compliance Architect / Product Security Sr. Engineer/Expert (f/m/d)
@ SAP | Walldorf, DE, 69190
SAP Security Administrator
@ FARO Technologies | EMEA-Portugal