Can someone explain the focus on weather & Paris - then answering the weather for New-York? Mistake or process?

Is there a simple explanation of the rlhf bit? How does it work and how does the feedback get "into" got?

Simple and somewhat decent intro to gpt in general. But I still feel this was somewhat repetitive.

I'm not that interested in LLMs as chat bots, especially not to "replace workers," but as mathematical models starting to be able to explain mid-level human behavior, they're really cool.

So you’re saying we can view them as mirrors or models that help us understand ourselves better.