Thank you.
#3
by
nanowell
- opened
I wanted to thank you for providing fine tuned model for Open source community, which beats gpt-4 in human eval benchmark. W Phind. This is the way.
I wanted to thank you for providing fine tuned model for Open source community, which beats gpt-4 in human eval benchmark. W Phind. This is the way.
You are correct, but only for the (pre)release gpt-4, gpt-4(and gpt3.5) now scores way higher, so this model does not beat the recent version of gpt-4(nor the recent version of gpt3.5). Still impressive though.