@goodside:

Idea: Using logit bias to adversarially suppress GPT-4’s preferred answers for directed exploration of its hallucinations.

Here, I ask: “Who are you?” but I suppress “AI language model”, “OpenAI”, etc.

This reliably elicits narratives about being made by Google:

(see screenshot in tweet, he also posted the code)

  • 𝕊𝕚𝕤𝕪𝕡𝕙𝕖𝕒𝕟@programming.devOPM
    link
    fedilink
    English
    arrow-up
    1
    ·
    1 year ago

    Thank you! I’m glad you like them!

    There’s so much noise and so little signal about AI out there that I think we really need a community focused on high quality content. Let’s hope it grows! I hope we can attract more people to this instance and the fediverse in general.