Really cool initiative, curious to see the results (is Llama 3.1 70B base the second-best judge LLM?)

→ View original post on X — @maximelabonne