TextSeal2: distortion-free LLM watermarking beats SynthID-Text and textGrain under speculative decoding
TimDarcet · x · 2026-11-02
TextSeal2 is a new distortion-free text watermarking method using copula mixing and memoryless races. Per the author's TLDR:
- Beats textGrain and Google's SynthID-Text under speculative decoding;
- Adds no extra latency, with the same acceptance rate;
- A single parameter fully controls the diversity/detectability tradeoff.
More from Research
- LLMs are getting good at directly writing neural network weights—what does that mean for gradient descent? — RichmanRonald · 2026-11-03
- UCSB awards IEE seed funding to three ECE projects on energy-efficient AI and quantum hardware — ninamiolane · 2026-11-03
- OpenAI deploys activation classifiers (probes) on frontier models starting from GPT-5.6 — bobabowen · 2026-11-03
- Google's ScientistTwo runs the full research loop autonomously, out-rating human papers — udmrzn · 2026-11-03
- ML4 tech report drops, detailing pretraining architecture and agentic performance interplay — Fluke_Ellington · 2026-11-03
- TuneJury: a model trained on human judgments predicts music preferences — mittu1204 · 2026-11-03