What retrieval strategies matter most to you?

[AmitoVrito]

We're building advanced retrieval beyond what LangChain offers. Which strategies should come first?

• Hybrid Search (BM25 + vector with RRF)
• GraphRAG (knowledge graph-based retrieval)
• RAPTOR (recursive abstractive processing)
• Self-RAG (self-reflective retrieval)
• CRAG (corrective RAG with web fallback)
• ColBERT (late interaction token-level matching)
• Contextual Retrieval (Anthropic-style chunk context)
• Cohere/Cross-Encoder reranking
• Parent Document retrieval
• RAG Fusion (multi-query with RRF)

Comment your top 3!