Interactive RAG Explainer
Learn the basics of Retrieval Augmented Generation (RAG) by interacting with each step of the pipeline.
1. Input Data
→
2. Indexing
→
3. Retrieval
→
4. Generation
2. Indexing
The document is broken down into smaller chunks (e.g., sentences or paragraphs). Hover over a segment in the 'Original Document' or an 'Index Entry' to see its counterpart highlighted.
Original Document (Interactive Segments)
Index Entries (Chunks)
4. Retrieval & Ranking
Index entries are retrieved and ranked based on their relevance to your question. The top-ranked chunks (e.g., top 3) form the 'context window' for the LLM.
5. Generation (Simulated)
The LLM uses the retrieved context to formulate an answer.