REVIEW 2 cited by
Debiasing Watermarks for Large Language Models via Maximal Coupling
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Watermarking language models is essential for distinguishing between human and machine-generated text and thus maintaining the integrity and trustworthiness of digital communication. We present a novel green/red list watermarking approach that partitions the token set into ``green'' and ``red'' lists, subtly increasing the generation probability for green tokens. To correct token distribution bias, our method employs maximal coupling, using a uniform coin flip to decide whether to apply bias correction, with the result embedded as a pseudorandom watermark signal. Theoretical analysis confirms this approach's unbiased nature and robust detection capabilities. Experimental results show that it outperforms prior techniques by preserving text quality while maintaining high detectability, and it demonstrates resilience to targeted modifications aimed at improving text quality. This research provides a promising watermarking solution for language models, balancing effective detection with minimal impact on text quality.
Forward citations
Cited by 2 Pith papers
-
Optimal Estimation of Watermark Proportions in Hybrid AI-Human Texts
For continuous-score text watermarks, the proportion of watermarked tokens in mixed AI-human text is identifiable and can be estimated at the minimax-optimal rate.
-
Watermark in the Classroom: A Conformal Framework for Adaptive AI Usage Detection
Standard, hierarchical, and weighted conformal prediction applied to LLM watermark scores can control false-positive rates when detecting guideline-violating AI edits in simulated classroom essays.
Discussion (0). Continue with ORCID to comment.