There’s some common LLM-powered workflows that users are orchestrating over their data: * Structured data extraction: input -> prompt --> LLM —> output parser * Naive RAG: input -> retriever -> synthesizer * Advanced RAG: input —> query rewriting —> retriever —> reranker —>… https://t.co/0bK0UTfn7B https://t.co/TedK7CSz11 https://t.co/Z6FO7Hs5M0
— Jerry Liu (@jerryjliu0) Jan 8, 2024
from Twitter https://twitter.com/jerryjliu0
January 08, 2024 at 05:15PM
via IFTTT
No comments:
Post a Comment