Yes, I think it's harder for humans to evaluate, but probably not for LLMs

→ View original post on X — @maximelabonne