WorldCup Sampling for Multi-bit LLM Watermarking

Yidan Wang, Yubing Ren, Yanan Cao, Li Guo · 2026

Read WorldCup Sampling for Multi-bit LLM Watermarking by Yidan Wang, Yubing Ren, Yanan Cao, Li Guo (2026) 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-2602-01752.txt.

Description

Yidan Wang and colleagues embed several bits in language model text, extending zero bit schemes beyond mere spotting. Methods extend zero bit schemes with fixed logit shifts and counting reads, which degrade text quality and read robustness at larger payloads. Their WorldCup framework models sampling as a structured channel, embedding bits through layered contests. Entropy aware tuning preserves quality, and confidence aware reads using token scores enable message recovery. Tests show a balance across capacity, robustness and quality, beating baselines and offering a base for future marking.

Catalog record

CREATORYidan Wang, Yubing Ren, Yanan Cao, Li Guo
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE115289
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602.01752

Browse the full library · About the Tower of Babel