Retrieval-Augmented Generation

/rag

Augments LLM generation with retrieved external knowledge to reduce hallucination and improve accuracy.

Strata: L2 Axes: Knowledge Status: seed

Explainer

RAG pipelines retrieve relevant documents from a knowledge base before generating a response, grounding the agent's output in real data rather than parametric memory alone. Implementations range from simple vector-search retrieval to multi-hop re-ranking pipelines with hybrid search.

Anti-Patterns

None documented yet.

Mitigations