Overview: Evaluate RAG quality
The old saying “you can’t manage what you can’t measure” is very relevant for any generative AI application, including RAG applications. For your generative AI application to deliver high-quality and accurate responses, you must be able to define and measure what “quality” means for your use case.
This section deep dives into three critical components of evaluation:
- Define “quality”: Evaluation sets
- Assess performance: Metrics that matter
- Enable measurement: Supporting infrastructure