Guess which of these LLM outputs is watermarked

(sgoedecke.github.io)

11 points | by gfysfm 8 hours ago

5 comments

  • Noumenon72 4 hours ago
    Please report success/failure after each test. Asking me to read and compare 30 writing samples to get any feedback at all means I won't finish. Telling me immediately when I got one wrong lets me recognize patterns and improve my guesses.
  • madarcho 5 hours ago
    If SynthID is a google technology, then this is likely just us training their ai again, captcha all over again.
  • NotPractical 6 hours ago
    Could do with some context on how watermarking works. Objectively speaking it should be impossible to tell.
  • arcwhite 8 hours ago
    Interesting, I did very badly, 3/10!
  • fwlr 6 hours ago
    Utterly imperceptible, even when studied under the microscope in a way that LLM text very rarely is in practice.

    It will be interesting to see whose concerns are assuaged (perhaps they genuinely though mistakenly believed it would degrade quality), and whose concerns are heightened (perhaps their real objection is that their AI-generated text will become detectable).