Sign in to confirm you’re not a bot
This helps protect our community. Learn more
Agentic RAG with reflection
Retrieval Augmented Generation (RAG) is a popular technique to get LLMs to answer questions based off your own knowledge base. "Agentic RAG" adds planning, tool use, and reflection to a RAG flow. In this session, we'll extend our most popular open-source RAG solution with a reflection step, prompting the LLM to realize whether it needs to repeat the retrieval in order to improve the answer quality. [eventID:25338]

Follow along using the transcript.

Microsoft Reactor

114K subscribers