This repository contains code for the paper On the Learnability of Watermarks for Language Models by Chenchen Gu, Xiang Lisa Li, Percy Liang, and Tatsunori Hashimoto. Additional documentation and experiments code and data will be released soon.
The watermarking code in the kgw_watermark
directory is from https://github.com/jwkirchenbauer/lm-watermarking, and the code in the kth
directory is from https://github.com/jthickstun/watermark.