Introducing Code Llama, a state-of-the-art large language model for coding

ijeff ( @ijeff@lemdro.id ) · 1 year ago

Introducing Code Llama, a state-of-the-art large language model for coding

d3Xt3r ( @d3Xt3r@beehaw.org ) · edit-2 1 year ago

Looks interesting, but doesn’t seem better than GPT-4. GPT-4 scored 67% on the Human Eval test, whereas Code Llama scored only a 53.7%, which isn’t a trivial difference. Bit disingenuous of Meta to claim it to be “on par” with ChatGPT.

ijeff ( @ijeff@lemdro.id ) · 1 year ago

They seem to qualify a bit below that they mean GPT-3.5-Turbo, which does often get referred to as ChatGPT (in contrast to GPT-4).