WaterSearch: A Quality-Aware Search-based Watermarking Framework for Large Language Models
Yukang Lin, Jiahao Shao, Shuoran Jiang, Wentao Zhu, Bingjie Lu, Xiangping Wu, Joanna Siebert, Qingcai Chen · 2025
Read WaterSearch: A Quality-Aware Search-based Watermarking Framework for Large Language Models by Yukang Lin, Jiahao Shao, Shuoran Jiang, Wentao Zhu, Bingjie Lu, Xiangping Wu, Joanna Siebert, Qingcai Chen (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2512-00837.txt.
Description
Yukang Lin and seven coauthors present WaterSearch in a 2025 paper on watermarking AI text. Watermarks hide a detectable signal inside text written by large language models. That signal helps prove where text came from, which supports accountability and trust.
Most methods steer token choices as text is generated, then detect the signal with a statistical test. Lin and coauthors take a search-based route that also weighs text quality. Their quality-aware framework looks for watermarked outputs that still read well. Balancing detection strength against readability is key if watermarks are to be used on real products.
Catalog record
| CREATOR | Yukang Lin, Jiahao Shao, Shuoran Jiang, Wentao Zhu, Bingjie Lu, Xiangping Wu, Joanna Siebert, Qingcai Chen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 61 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-00837