SynthID-Text draws several candidate tokens from the model's unmodified distribution and compares them through keyed tournament rounds. Green-list decisions select winners without applying one fixed bias directly to every preferred token logit.
Detection repeats the keyed verdict process across tokens and rounds, then aggregates the evidence. The design aims to preserve response quality, but detectability still depends on output length, token entropy, key access, and whether later editing or paraphrasing disrupts the signal.
Acronyms and aliases
Google SynthID-Text variant
General terms
Related terms
Frequently asked questions
How does SynthID-Text differ from fixed logit-bias watermarking?
It selects among honest samples with keyed tournaments rather than adding one constant bias directly to preferred token logits.
Does SynthID-Text make an artificial intelligence watermark universally visible?
No. Detection requires compatible keyed analysis, and short, low-entropy, or paraphrased output can weaken the signal.
