Hi,Could you please share the commands you used to run evals on HumanEval and MBPP tasks so I can try to reproduce the reported numbers?I am not sure if I am hitting the right params like temp, top_p, etc.
· Sign up or log in to comment