Hybrid RAG & Re-ranking Lab

Compare Dense Vector Search vs Sparse BM25 vs Hybrid Reciprocal Rank Fusion (RRF) with Cross-Encoder Re-Ranking.

RAG Pipeline Configuration

Retrieval Strategy:

Chunk Size:256 tokens

User Query:

Retrieved & Re-Ranked Document Chunks0 Chunks Matched

Click “Run RAG Retrieval” to execute search across the corpus.

Loading AI::COSMOS...