[riku_iki]

I think it is valid criticism that that HumanEval benchmark is not completely representative, they also say it in the post.