Harrison Chase：用多个维度评估智能体运行轨迹
中文全文 · AI 翻译
轨迹标注实际上要回答好几个问题，而不只是给出一个通过或失败的判断——这里的阐述方式很不错。
这正是 Jev 在 LangSmith 评估中担任评判者的工作方式：对每条运行轨迹，只需一轮处理，就能分别为难度和正确性给出各自带有明确类型的答案。
https://www.langchain.com/blog/jev-is-now-available-in-langsmith-evals
对照原文
trajectory labeling is really several questions, not one pass/fail - nice framing here that's how Jev as a judge works in LangSmith evals: difficulty and correctness each get their own typed answer, in one pass, on every trace https://t.co/jlD2ywtlOT https://t.co/QgQpQFpZ26