Comment on Words Without Consequence

rimu@piefed.social ⁨3⁩ ⁨weeks⁩ ago

• Powderhorn@beehaw.org ⁨3⁩ ⁨weeks⁩ ago

  What, you saw a couple of em-dashes and reached that conclusion? I would argue – emphatically – that people are using the wrong methods to attempt to unmask LLMs. They’re literally trained on prose written by humans. We still know how to write.

  source

  • rimu@piefed.social ⁨3⁩ ⁨weeks⁩ ago

    No.

    Copy and paste the body of your post into https://copyleaks.com/ai-content-detector or https://gptzero.me.

    It’s ok to be deceived, I don’t blame you for it.

    source

    • Hexarei@beehaw.org ⁨3⁩ ⁨weeks⁩ ago

      Those AI detectors don’t work, btw.

      source
    • Powderhorn@beehaw.org ⁨3⁩ ⁨weeks⁩ ago

      So that I fully understand your argument, it is that other automation proves that what I posted is LLM output? Let’s sit with that for a moment. Your bullshit detector is accurate, but mine is not. This is essentially your thesis.

      I’d workshop this for a bit.

      source

      • rimu@piefed.social ⁨3⁩ ⁨weeks⁩ ago

        Ohhh, I get it now. Thanks for spelling it out.

        Sorry for hurting your feelings. I didn’t mean to imply your bs detector is broken but I can see how it would come across that way.

        source

        • -> View More Comments
• hono4kami@piefed.social ⁨3⁩ ⁨weeks⁩ ago

  What made you think it’s generated by LLM?

  source