The Kirchenbauer et al. (2023) watermark nudges an LLM to prefer a pseudo-random green list of tokens at each step, chosen by hashing the previous token. Normal text uses green tokens about a fraction γ of the time. Watermarked text uses far more, which a z-test detects.
Unlock this and every premium problem across all tracks — plus hidden-test grading, advanced AI engineering modules and new content every week.
Start 7-day free trial