• 0 Posts
  • 6 Comments
Joined 1 year ago
cake
Cake day: June 12th, 2023

help-circle


  • Best is very subjective.

    .world is a good general purpose instance for just about anything. I think it has the biggest population at the moment, so communities there are likely to get at least some engagement.

    For “general discussion” it doesn’t really matter. The instances are federated so you’ll likely get general discussion in comments from lots of people from lots of instances anyway, wherever your community is based.

    Some people get almost nationalistic about their chosen instances or have grudges against people from certain other instances. There’s sometimes inter-instance politics with some servers defederating with others or threatening to for various reasons. It’s kinda fun to watch in a popcorn drama kind of way. For the most part, the instance doesn’t matter.




  • I think the idea is that there are potentially alignment issues in LLMs because it’s not clear what concepts map to what activations. That makes it difficult to see what they’re really “thinking” about when they generate text. Eg. if they’re being misleading or are incorrectly associating concepts that shouldn’t be connected etc.

    The idea here is to use some mechanistic interpretability stuff to see what text activates what neurons in an LLM and then crowd source the meanings behind that and see if that’s something you could use to look up some context from an ai. Sort of trying to make a “Wikipedia of AI mind reading”

    Dunno how practical it is or how effective that approach is but it’s an interesting idea.