mathling, 10 months ago #UO: RAG (retrieval augmented generation) Prep: documents, maybe markup-rich and create embeddings from them (vectors) stored in vector database Use: create embeddings of question; similarity search of that vector from DB; augment prompt with those retrievals #balisage
#UO: RAG (retrieval augmented generation)
Prep: documents, maybe markup-rich and create embeddings from them (vectors) stored in vector database
Use: create embeddings of question; similarity search of that vector from DB; augment prompt with those retrievals
#balisage