Paper Detail

TextSeal2: Distortion-Free LLM Watermarking with Copula Mixing and Memoryless Races

Tom Sander, Pierre Fernandez, Itai Gat, Haiming Bao, Jiaxin Shi, Jason Park, Valeriu Lacatusu, Sylvestre-Alvise Rebuffi, Hongyan Chang, Hady Elsahar, Tomáš Souček, Tuan Tran, Surya Parimi, Alexandre Mourachko

arxiv Score 11.3

Published 2026-10-29 · First seen 2026-10-30

General AI

Abstract

We present TextSeal2, a distortion-free watermark for large language models that keeps its signal under speculative decoding: tokens that replace rejected drafts remain watermarked with the same keys, and the acceptance rate is unchanged. It also preserves output diversity and adds no measurable inference latency. Its predecessor, TextSeal, restores diversity in the Gumbel-max scheme by routing each token to one of two keys via a coin flip. This is suboptimal because the detector averages the scores from both keys without knowing which key was used. TextSeal2 instead merges two or more keys into a single virtual key and couples it to fresh noise through AutoMark's Gaussian copula. Any number of keys, including publicly shared keys, can be merged this way without diluting detection. Under speculative decoding, TextSeal2 verifies keyed drafts with one more exponential race, run on the leftovers of the draft races. By memorylessness, this race reproduces block verification, so the output follows the target model exactly. We evaluate TextSeal2 on three LLMs. At equal output token diversity, TextSeal2 detects 8 to 22 percentage points more 100-token answers than TextSeal at a false-positive rate of $10^{-3}$, and is comparable to textGrain while requiring no post-logits optimization. Under speculative decoding, it detects 18 and 22 percentage points more 200-token answers than TextSeal and textGrain, respectively, at the same acceptance rate.

Workflow Status

Review status
pending
Role
unreviewed
Read priority
now
Vote
Not set.
Saved
no
Collections
Not filed yet.
Next action
Not filled yet.

Reading Brief

No structured notes yet. Add `summary_sections`, `why_relevant`, `claim_impact`, or `next_action` in `papers.jsonl` to enrich this view.

Why It Surfaced

No ranking explanation is available yet.

Tags

No tags.

BibTeX

@article{sander2026textseal2,
  title = {TextSeal2: Distortion-Free LLM Watermarking with Copula Mixing and Memoryless Races},
  author = {Tom Sander and Pierre Fernandez and Itai Gat and Haiming Bao and Jiaxin Shi and Jason Park and Valeriu Lacatusu and Sylvestre-Alvise Rebuffi and Hongyan Chang and Hady Elsahar and Tomáš Souček and Tuan Tran and Surya Parimi and Alexandre Mourachko},
  year = {2026},
  abstract = {We present TextSeal2, a distortion-free watermark for large language models that keeps its signal under speculative decoding: tokens that replace rejected drafts remain watermarked with the same keys, and the acceptance rate is unchanged. It also preserves output diversity and adds no measurable inference latency. Its predecessor, TextSeal, restores diversity in the Gumbel-max scheme by routing each token to one of two keys via a coin flip. This is suboptimal because the detector averages the sc},
  url = {https://arxiv.org/abs/2610.40613},
  keywords = {cs.CR, cs.AI, cs.CL},
  eprint = {2610.40613},
  archiveprefix = {arXiv},
}

Metadata

{}