Add Root Cause Analysis to Evaluation Runner/ELI5

The Root Cause Analysis (/troubleshoot) can currently only be executed through promptlib. Adding the dataset to ELI5 would enable us to run the eval for groupcustom models via the Evaluation Runner.

Dataset

LLM Judge

Definition of Done

  • one can run the / troubleshoot validation dataset in Validation Runner
  • Both the dataset and the evaluator should match what is in promptlib.
Edited by Susie Bitters