Code

I’m using mistral-7b to understand LLMs’ procedure.

Does anyone have an idea to improve this process?

do not recommend changing the number of tokens -> 1. :)