Re:Things we learned about LLMs in 2024
by zmgsabst on Wednesday January 01, 2025 @03:06AM
(#42564656)
Why shouldn’t they be able to do this?
DNNs implicitly learn a type theory, which they then reason in. Even though the code itself is new, it’s expressible in the learned theory — so the DNN can operate on it.
DNNs implicitly learn a type theory, which they then reason in. Even though the code itself is new, it’s expressible in the learned theory — so the DNN can operate on it.