RAG
Recall@k
Recall@k is the fraction of relevant items found among the top k results.
How it works
It measures whether retrieval surfaces the needed evidence before generation. Increasing k cannot reduce recall for the same ranking, but it can add distracting context.
A concrete example
Illustrative: 4 of 5 relevant passages appear in the top 10, so Recall@10 = 0.80.
Do not confuse it with Precision@k
recall asks how much relevant material was found; precision asks how much retrieved material was relevant.