alt Hacker NewsNo one is claiming that every sentence LLMs are producing are literal copies of other sentences. Tokens are not even constrained to words but consist of smaller slices, comparable to syllables. Which even makes new words totally possible.
New sentences, words, or whatever is entirely possible, and yes, repeating a string (especially if you prompt it) is entirely possible, and not surprising at all. But all that comes from trained data, predicting the most probably next "syllable". It will never leave that realm, because it's not able to. It's like approaching an Italian who has never learned or heard any other language to speak French. It can't.
Replies
Your view of what is happening in the neural net of an LLM is too simplistic. They likely aren't subject to any constraints that humans aren't also in the regard you are describing. What I do know to be true is that they have internalised mechanisms for non-verbalised reasoning. I see proof of this every day when I use the frontier models at work.
> It's like approaching an Italian who has never learned or heard any other language to speak French
Interesting similitude, because I expect an Italian to be able to communicate somewhat successfully with a French person (and vice versa) even if they do not share a language.
The two languages are likely fairly similar in latent space.