Abstract:Watermarking acts as a critical safeguard in text generated by Large Language Models (LLMs). By embedding identifiable signals into model outputs, watermarking enables reliable attribution and enhances the security of machine-generated content. Existing approaches typically embed signals by manipulating token generation probabilities. Despite their effectiveness, these methods inherently face a trade-off between detectability and text quality: the signal strength and randomness required for robust watermarking tend to degrade the performance of downstream tasks.
In this paper, we design a novel embedding scheme that controls seed pools to facilitate diverse parallel generation of watermarked text. Based on that scheme, we propose WaterSearch, a sentence-level, search-based watermarking framework adaptable to a wide range of existing methods. WaterSearch enhances text quality by jointly optimizing two key aspects: 1) distribution fidelity and 2) watermark signal characteristics. Furthermore, WaterSearch is complemented by a sentence-level detection method with strong attack robustness. We evaluate our method on three popular LLMs across ten diverse tasks. Extensive experiments demonstrate that our method achieves an average performance improvement of 51.01\% over state-of-the-art baselines at a watermark detectability strength of 95\%. In challenging scenarios such as short text generation and low-entropy output generation, our method yields performance gains of 47.78\% and 36.47\%, respectively. Moreover, under different attack senarios including insertion, synonym substitution and paraphrase attasks, WaterSearch maintains high detectability, further validating its robust anti-attack capabilities. Our code is available at \href{this https URL}{this https URL}.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2512.00837 [cs.CL] |
| (or arXiv:2512.00837v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2512.00837 arXiv-issued DOI via DataCite |
Submission history
From: Yukang Lin [view email]
[v1]
Sun, 30 Nov 2025 11:11:21 UTC (281 KB)
[v2]
Thu, 28 May 2026 09:46:18 UTC (275 KB)
[v3]
Sun, 31 May 2026 07:46:22 UTC (275 KB)