LAB · RAG
Retrieval Metrics Lab
Calculate precision@k, recall@k, reciprocal rank, and nDCG from labeled relevance.
- Suggested time
- 12 min
- Output artifact
- A report that separates retrieval failure from generation failure.
- Privacy
- Only identifiers and relevance labels entered locally are used.
Local educational tool. Validate every result before production use.
When to use it
Retrieval metrics
Use it when you can label relevant documents per query and need to compare retrievers or values of k.
How to use it well
- 1
Enter a real case without secrets or personal data.
- 2
Run, inspect, and challenge the assumptions; the output is not automatic truth.
- 3
Save the artifact in your project and turn it into a reviewable test or decision.
Related foundations
Retrieval Engineering
07 — Evaluating RAG with RAGAS
A RAG system has too many knobs (chunking, k, embedding model, hybrid yes/no, reranker, prompt...) to tune by eye. Without automated evaluation, every change is a gamble: it might improve the 5 questions you test manuall
Open foundation20 minLab 03 — Compare chunking with a retrieval metric
Localized walkthrough. Run the canonical public lab locally and preserve the result as evidence.
Open foundation20 minLab 06 — Layered evaluation: deterministic retrieval and RAGAS 0.4.x
Localized walkthrough. Run the canonical public lab locally and preserve the result as evidence.
Open foundation