Skip to content

RAGAS

πŸ“ Short definition

Framework that evaluates a RAG system by separating retrieval quality and fidelity of the generated answer.

πŸ“– Detailed definition

RAGAS measures in particular:

  • faithfulness: is the answer supported by the retrieved contexts?
  • answer relevancy: does the answer actually address the question?
  • context precision: are the retrieved passages relevant?
  • context recall: does retrieval find the necessary information?

The value is diagnosing where the system fails: poor document retrieval, good source poorly used, off-topic answer, or hallucination.

πŸ’‘ Why it matters for on-prem AI

Many local assistants rely on RAG to use internal documents. RAGAS helps measure that layer’s quality without relying only on subjective impression.

⚠️ Common pitfalls

  • Believing a good model compensates for poor retrieval.
  • Measuring only the final answer without checking retrieved passages.
  • Replacing all human review with an automatic score.

πŸ”— See also

Give my opinion

Everything is optional: one answer is enough. This site sends nothing: you send it, the way you choose.

Your noteNo note
Your opinion is about
About what? (several choices possible)
What will be sent
Send by e-mail

Private, no account: opens your mail program with the message ready, for ia-on-prem@damien.becherini.fr.

Open a GitHub issue β†—

Public, with a GitHub account. You read the issue again before you publish it.

To paste it wherever you want.

Give at least a note, a subject or a message to be able to send.