raganwald, ChatGPT 3.5 vs. ChatGPT4.
Question: Is there something radically improved in the ChatGPT4 LLM such that it gets this right "organically?"
Or—pardon my cynicism—is there some exception code slapped into the system in response to a bug being filed, and this isn't actually "improved" so much as "faked to avoid embarrassment?"
Add comment