LLMs work differently, statistically predicting the next token (roughly equivalent to a word) based on all those that came before it, and parameters finetuned during training.
Which is what a parrot does.
Yeah this is the exact criticism. They recombine language pieces without really doing language. The end result looks like language, but it lacks any of the important characteristics of language such as meaning and intention.
If I say “Two plus two is four” I am communicating my belief about mathematics.
If an llm emits “two plus two is four” it is outputting a stochastically selected series of tokens linked by probabilities derived from training data. If the statement is true or false then that is accidental.
Hence, stochastic parrot.
If i train an LLM to do math, for the training data i generate a+b=cstatements, never showing it the same one twice.
It would be pointless for it to “memorize” every single question and answer it gets since it would never see that question again. The only way it would be able to generate correct answers would be if it gained a concept of what numbers are, and how the add operation operates on them to create a new number.
Rather than memorizing and parroting it would have to actually understand it in order to generate responses.
It’s called generalization, it’s why large amounts of data is required (if you show the same data again and again then memorizing becomes a viable strategy)
If I say “Two plus two is four” I am communicating my belief about mathematics.
Seems like a pointless distinction, you were told it so you believe it to be the case? Why can’t we say the LLM outputs what it believes is the correct answer? You’re both just making some statement based on your prior experiences which may or may not be true
You’re arguing against a position I didn’t put forward. Also
Seems like a pointless distinction, you were told it so you believe it to be the case? Why can’t we say the LLM outputs what it believes is the correct answer? You’re both just making some statement based on your prior experiences which may or may not be true
This is what excessive reduction does to a mfer. That is just such a hysterically absurd take.
If you fine tune a LLM on math equations, odds are it won’t actually learn how to reliably solve novel problems. Just the same as it won’t become a subject matter expert on any topic, but it’s a lot harder to write simple math that “looks, but is not, correct” than it is to waffle vaguely about a topic. The idea of a LLM creating a robust model of the semantics of the text it’s trained on is, at face value, plausible; it just doesn’t seem to actually happen in practice.
You take in some information, combine that with some precious experiences, and then output words
Which is what an LLM does.
Flat epistemological statements like this are why I feel like more STEM people need to take Philosophy.
