- RAG Architecture · Generative AI
LLearnViaNotes @learnvianotes· 1 min read Retrieval-Augmented Generation grounds LLM answers in external source documentation.
Complete RAG Pipeline Flow
1. Parse Docs PDFs, HTML, Markdown → 2. Chunking Size: 512, Overlap: 50 → 3. Embed Text -> 1536-dim vector → 4. Store Vector DB Index
5. User Query "How to refund?" → 6. Retrieval Vector Cosine Search → 7. Context Inject Format prompt with facts → 8. LLM Answer Fact-grounded response
Free for students and teachers
Write notes like this one
Cards for every idea, quizzes and flashcards, read aloud — on the web and your phone.
Something wrong with this note? Open it to report