Pass metadata as parameter into LLM-as-a-judge config in langsmith UI

Product Help LangSmith
1

Hi there!

I have a question. Is it possible to pass metadata into LLM-as-a-judge evaluator?
This is about online and over-a-dataset evaluators.

I know it is possible with SDK, but is there any way to do it in langsmith UI?

Thanks!

2

No – currently you can only pull from example inputs and reference outputs when using LLM-as-a-judge evaluators in the LangSmith UI.