Anthropic's watermarking uses token sampling statistics, not hidden chars
Scobleizer · x · 2026-08-15
Anthropic's watermarking technique modifies the source of randomness for token selection rather than inserting hidden characters. Based on Google DeepMind's SynthID-Text, it introduces a key or context-derived sampling step within the probability cloud. While individual choices are undetectable, hundreds or thousands of tokens accumulate into a recognizable statistical signature for key holders. The output itself acts as the watermark.
Related event: Anthropic Deploys Text Watermarking for Claude to Comply with EU AI Act(31 posts)→
More from Safety
- Linux Kernel CVEs Surge From ~500 to 1500+ Per Release, LLMs Blamed for Bulk of the Rise — burny_tech · 2026-10-03
- COLM 2026 Launches DAIH Workshop on Deploying LLMs/VLMs Responsibly in Healthcare — StellaLisy · 2026-10-03
- Trillium Labs wants to do open research on recursive self-improvement and agents — nordicinst · 2026-10-03
- Trillium Labs Wants to Research Self-Improvement and Model Behavior in the Open — Wired AI · 2026-10-03
- Cloudflare Turnstile everywhere: anti-AI scraping walls now hit human users — sethlazar · 2026-10-02
- Filler tokens let frontier models reason invisibly: 13-point gains undetectable by CoT monitoring — PandaAshwinee · 2026-10-02