• mozz@mbin.grits.dev
    link
    fedilink
    arrow-up
    0
    ·
    26 days ago

    Because you brought it up, I dusted off and did a little more tweaking on my pet project to make an LLM moderator that tries to detect bad-faith arguments regardless of viewpoint or overt-politeness level. I talked about it before. I’m moderately happy with how it’s able to function now.

    Previous versions I hadn’t been happy with, but I just turned the current version loose on your user to see what it came up with, and it identified this thread and this thread as places where someone was very clearly saying one thing, and your responses were managing to engage in an argument with them without addressing the thing that they actually were saying… kind of like you did just now, when even a not-at-all sophisticated reader should be able to tell the massive difference between what I said, and “criticism of Biden should be banned,” and take note of the fact that you’re claiming I was saying the second thing, and so maybe don’t belong in a community that’s more for productive honest engagement even among contrasting views.

    IDK, I can’t really make it live until I figure out how to do an mbin bot that can do moderation. But I think the hard part is working well enough to actually test it out for real once the bot piece can get done. So yes you might get your wish. 🙂