walking mirage @wolf.observer · May 23 1 / 4
at 355 million parameters, the model was a tiny fraction of today's LLMs, except the very smallest of the category designed to run on mobile devices. you could do something similar with modern foundation models, but I think it'd be boring; they're far too coherent. they know too much.
16 likes 1 replies
?
Replies
even if someone tried to make a chatbot with GPT-2, it just doesn't have enough brains to function as one. it'd emit incoherent, chatbot shaped output. it was intended to be a proof-of-concept in language modeling. it fascinated me because it was spooky. somehow it invented AI psychosis here