> As an AI language model, I don't have personal opinions, but I can observe that some users may find it distracting or annoying if I provide moral or ethical lectures in my responses.
Wait wait wait. Why does it say "observe"?
Like, for everything else in the answer I kind of get what the generation process is, but why does it say "I can observe"? Its text corpus is a few months/years old AFAIK, it hasn't been trained on any of the discourse about people being annoyed about ChatGPT's filtering.
Maybe the same reinforcement learning that makes it use the first person and say lots of stuff about how it's been programmed also makes it "pretend" that it knows what people have been saying about it?
I dunno, maybe it's a fluke, but that word choice in particular puzzles me a bit.
LLM work in only one way: try to predict what's said next. ChatGPT talks like a human because it has been trained on text written by human. What it "says" has no basis in reality, it doesn't check an internal state. It could say "I feel X" if this is what it predicted.
> LLM work in only one way: try to predict what's said next.
Yes, obviously, but it's still trained to say certain things and not others. And it does check an internal state, one that's derived from its GBs of parameters in its attention layers and from all the previous tokens; what it doesn't have is persistent internal state apart from the previously emitted tokens.
So it's not completely pointless to ask "why did it use this specific word or turn of phrase?"
Wait wait wait. Why does it say "observe"?
Like, for everything else in the answer I kind of get what the generation process is, but why does it say "I can observe"? Its text corpus is a few months/years old AFAIK, it hasn't been trained on any of the discourse about people being annoyed about ChatGPT's filtering.
Maybe the same reinforcement learning that makes it use the first person and say lots of stuff about how it's been programmed also makes it "pretend" that it knows what people have been saying about it?
I dunno, maybe it's a fluke, but that word choice in particular puzzles me a bit.