Retrieval and RAG
Context recall
Implemented
rag.context_recallDefinition
Share of the reference answer's claims that can be attributed to the retrieved context (RAGAS context recall): did retrieval find the evidence needed?
Formula
reference claims supported by the context / reference claims
Range: [0, 1]
Inputs and outputs
- claim_attributed: per query, whether each reference claim is in the context
Returns: MetricResult (float, or per-example array with average=None)
Assumptions
No assumptions beyond valid, aligned inputs of the documented types.
Limitations
No metric-specific limitations are documented yet. Interpret the value alongside the task, data, and other metrics.
Python API
import evalsuite as es
es.context_recall(claim_attributed)References
- Es S, James J, Espinosa-Anke L, Schockaert S. RAGAS: automated evaluation of retrieval augmented generation. EACL (demos). 2024:150-158.
- Ru D, Qiu L, Hu X, et al. RAGChecker: a fine-grained framework for diagnosing retrieval-augmented generation. NeurIPS Datasets and Benchmarks. 2024.