feat(Datasets): Filter Dataset runs by metadata / tags

[Omri-Gold]

Describe the feature or potential improvement

I’d like the ability to filter dataset runs by metadata or by tags. I know tags for dataset runs are not yet available but are planned, and this feature would complement that roadmap.

We use dataset runs in our CI as part of a sanity test process. These CI runs currently get mixed together with our real evaluation runs, which pollutes the score/latency graphs. Being able to filter runs - either by metadata or by tags - would allow us to cleanly separate CI runs from actual evaluation runs. This becomes even more important as we move towards supporting multiple dataset versions.

Additional information

No response

[marliessophie]

Hey, thanks for suggesting this. Agree this is a useful feature. I will keep you posted here if we start working on either.