What is RAG
RAG is an approach where the model first retrieves relevant fragments from your knowledge base and grounds its answer on them.
Architecture
- Ingest and chunk documents
- Vectorize (embeddings)
- Similarity search
- Generate an answer with citations