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

• root cause analysis (/troubleshoot)

LLM Judge

• LLM Judge rca_foundational_models/judge-system.txt

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.