Get the latest tech news
Questions about LLMs in Group Chats
Lately, I’ve been keeping a close eye on the xenocog community and projects that are spinning out of the ideas of the cyborgism and act 1 projects. However, in classic fashion, the aspects that are intriguing me the most are not what most people are interested in.
Either way I’m sure I’m not the first one to think about this and wanted to look into what other methods people are taking and create a more definitive list of tunable features before I set out to make my own framework. In the settings of a shape the free will section has its own selection for a LLM model distinct from the regular flow, making me think it’s handling it separately on the backend with some kind of randomness. Some further analysis of the architecture of ai-town may lead to more interesting questions, but I don’t want to oversubscribe on this method since they were going for a very specific use case that doesn’t perfectly match group chats.
Or read this on Hacker News