Metraipolis › archive › dc_dd5c91c79f90
TTMark: Pairwise Distortion-Free Watermarking Beyond Single-Token Entropy — arXiv:2609.36372
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.36372
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T16:57:24.762Z
- SHA-256 of the archived text
fe4b4b8a542ed42d485f55f99a09f007c1d6583ab4669ba4095fce00e5ff5ee4- Stable URL of this page
https://metraipolis.com/documento/dc_dd5c91c79f90
How to cite
TTMark: Pairwise Distortion-Free Watermarking Beyond Single-Token Entropy — arXiv:2609.36372 — arXiv (en), https://arxiv.org/abs/2609.36372 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_dd5c91c79f90 (Metraipolis archive dc_dd5c91c79f90, sha256 fe4b4b8a542ed42d485f55f99a09f007c1d6583ab4669ba4095fce00e5ff5ee4)
BibTeX
@misc{dcdd5c91c79f90,
title = {TTMark: Pairwise Distortion-Free Watermarking Beyond Single-Token Entropy — arXiv:2609.36372},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.36372},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_dd5c91c79f90; sha256 fe4b4b8a542ed42d485f55f99a09f007c1d6583ab4669ba4095fce00e5ff5ee4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.36372. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Distortion-free watermarking enables reliable attribution of machine-generated text while preserving output distribution. However, existing methods operate independently on each generated token, making their detection capability fundamentally constrained by the entropy of the next-token distribution. We present Tandem Token WaterMark (TTMARK), a general pairwise watermarking framework that extends distortion-free watermarking from individual tokens to adjacent token pairs. By watermarking the joint distribution of consecutive tokens, TTMARK enlarges the effective watermarking alphabet from V to $V^2$, allowing the detector to exploit both token entropy and conditional entropy while preserving distortion-freeness over the joint distribution. We further introduce a branch-isolating concatenated tandem generation algorithm that efficiently constructs the joint distribution in a single forward pass. Theoretically, we show that pairwise watermarking achieves better expected detection strength in low-entropy regimes. Extensive experiments across multiple language models, datasets, and three representative distortion-free watermarking schemes demonstrate that TTMARK consistently improves detectability without degrading generation quality, while also improving robustness to edits and substantially enhancing localized watermark detection. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-30 in cs.CL). Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2609.36372 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T16:27:03.394Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-30?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).