1
LLM-as-a-Judge: I Built One From Scratch, Then Checked It Against Humans
从零搭建LLM裁判系统，与人类评分对比，揭露看似好用实则暗藏陷阱的评估局限。
In Part 1 the model's job was to pick one of 77 labels, so I could check it with == . But most real LLM output isn't like that — it's a paragraph, a s…