Remix.run Logo
Herring 4 days ago

I have an idea. What if we used a third LLM to evaluate how good the secondary LLM is at critiquing the primary LLM.