Separate model quality from retrieval quality: answer quality fell. Which metric separates the two stages?

prerequisite: GENEVAL-004

Context fixture

{
  "metric": "retrieval recall vs answer groundedness"
}

Hints · 0/3

Your answer

Opening…