Experiments Bug(?) - Every output is using the same input

Is anyone else running into issues with Langsmith experiments tab in their UI?

Every single one of my outputs seems to be referencing the same input so I think somehow it’s only pulling the first input to run all the tests?

Hey @promptlordjess,

I’m unable to repro - can you share more details on your dataset and exactly how you’re running things?

Yes this is what I do.

1 I click datasets&experiemnts tab
2 I open my golden labels data set
3. I click +Experiment → Run in Playground
4. I select my prompt, my 5 examples from that golden labels dataset, then I pick my evaluator
5. I hit run
6. I look through the output column and realize every thing was run against one of the examples instead of 5 different examples

Here’s the link if it helps: LangSmith