Advanced RAG Techniques.md

created	modified	tags	type	status
2024-10-01T20:49	2024-10-16 12:55	rag llm large-language-model llm-agents llama llama-index langchain	map-of-content	ongoing

Query Expansion/Augmentation

Technique	Explanation	Link(s)
Hallucinated answers	LLM is used to hallucinate an answer to the user query, and this hallucinated answer is combined with the original user query in the vector search	Youtube: RAG Fundamentals and Advanced Techniques – Full Course (free code camp)
Self-querying retrieval	LLM is prompted to use a querying interface in order to fetch data (e.g. a specific JSON format, SQL etc.)

Embeddings

Technique	Explanation	Link(s)

Reranking

Technique	Explanation	Link(s)

References

• Youtube: RAG Fundamentals and Advanced Techniques – Full Course (free code camp)

Related

• [[Graph RAG]]