Hybrid vs. Dense Retrievers in Blended RAG: Recall Performance on TriviaQA Across Corpus Sizes

This report synthesises findings from 12 peer-reviewed papers addressing the following research question: How does the retrieval recall of hybrid retrievers in Blended RAG compare to dense-only retrievers on the TriviaQA benchmark when evaluated across varying corpus sizes. 7 claims were extracted from source literature; 7 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 8.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the retrieval recall of hybrid retrievers in Blended RAG compare to dense-only retrievers on the TriviaQA benchmark when evaluated across varying corpus sizes?

Autonomous literature synthesis. Automated review score: 8.2/10. Full text and citation available at Assignee Research.