#llm-as-judge

I built an LLM-as-judge eval on my own blog and got a suspiciously perfect 16/16. Here's the three-round test I ran before trusting that number: single-variable corruption, and a self-consistency check the research says most teams skip.