#RAG

2 articles tagged RAG.

• Vector Search Beyond Embeddings: Hybrid Dense-Sparse Retrieval and Graph Traversal Foundation Models • 2026-10-29 • Dr. Elena Rostova

  Why pure vector embeddings fail on exact code identifiers, and how Reciprocal Rank Fusion combining BM25, dense vectors, and AST graph traversal achieves 94%+ retrieval recall.

• Fine-Tuning vs RAG vs In-Context Learning: An Empirical Decision Matrix Foundation Models • 2026-10-29 • Dr. Elena Rostova

  An empirical engineering framework for choosing between RAG, LoRA fine-tuning, and in-context learning based on data velocity, hallucination tolerance, and inference cost.