AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
OpenAI's Ability to Understand a Long Complicated Task in Sequence
I was interested in capabilities that were just on the fringe of reliability that nobody had really thought to chase yet. And after enough tuning on these examples, it somehow generalized that it got better at following instructions of a previously unseen length. I could tell that those were the sort of the prompts worth considering now because models are going to improve. RLHF marches onward: These models become more capable and more able to follow more complicated directions reliably.