• nous
    link
    fedilink
    English
    arrow-up
    12
    ·
    7 months ago

    How do we know these are the AI chatbots instructions and not just instructions it made up? They make things up all the time, why do we trust it in this instance?

      • neoman4426@fedia.io
        link
        fedilink
        arrow-up
        5
        ·
        7 months ago

        I tried it a few days ago and got some variance … but it was still exactly the same essential instructions, just a first person summary rather than the second person verbatim

      • DarkThoughts@fedia.io
        link
        fedilink
        arrow-up
        1
        ·
        7 months ago

        The cool thing with models who can do this is that you can kinda talk to the LLM behind whatever it is supposed to represent & change things dynamically (with respect to its context size of course). Not all models can do that unfortunately.