Hi 馃槉
Deltacoder is a nice LLM and I like it to write simple functions. However, I noticed that there is a serious error: It absolutely insists, that a function ... as task must have early returns and hallucinates reasons, why this has to be so. Moreover: It does not rewrite the function to use exactly one return at the end even if directly ordered to - and then tells the user, that is has followed the order.
Anthropomorphising: It's a stubborn idiot 馃ぃ
Hi 馃槉
Deltacoder is a nice LLM and I like it to write simple functions. However, I noticed that there is a serious error: It absolutely insists, that a function ... as task must have early returns and hallucinates reasons, why this has to be so. Moreover: It does not rewrite the function to use exactly one return at the end even if directly ordered to - and then tells the user, that is has followed the order.
Anthropomorphising: It's a stubborn idiot 馃ぃ