Evaluations

Faithfulness, retrieval relevance and hallucination scores are computed per run. Open any run in the Trace Explorer to see its scorecard, or re-run evaluation there.

Failures by agent
Failed runs (0)
No failures 🎉