Beyond embeddings: Navigating the shift to completions-only RAG | Sensible Blog
We’ve recently explored some new approaches to retrieval-augmented generation (RAG) that rely solely on completions without using embeddings. Learn how this completions-only method compares to embedding-based approaches, and why we believe it may be the future for certain RAG use cases as language models continue to improve. Strict timelines for technological advancement are inevitably inaccurate, yet progress occurs nonetheless. Businesses need AI maturity models, not timelines, to navigate the profound industry changes that will result as AI technologies become normalized. Extract structured data from financial documents in seconds Automate data entry from loss runs, ACORD forms, policies, and more Parse structured data from offering memos, rent rolls, and more Real-time bank statement processing Extract identification details from driver's licenses Instantly parse policy declaration pages Extract data from utility bills in seconds Bringing structure to unstructured data Companies us
A brief history of embeddings Language is just numbers. It's a statement that's not too hard to believe given the rapid progress of LLMs over the past few years. If you're a computer scientist you salivate at the prospect: all of semantics reduced to orderly vectors, the central primitive of decades of machine learning and statistics research. A notable LLM precursor, Word2vec, was the first to demonstrate the power of semantics-as-vectors. Word2vec’s goal is to represent the meaning of individual words as numeric vectors such that numeric operations on the vectors are semantically…
saved by
related reading
- Patterns for Building LLM-based Systems & Productseugeneyan.com
- Better RAG 3: The text is your friendolickel.com
- Better RAG 1: Basicsolickel.com
- Embeddings: What they are and why they mattersimonwillison.net
- Vinija's Notes • NLP • Retrieval Augmented Generationvinija.ai
- Retrieval-Augmented Generation for Large Language Models: A Surveyarxiv.org
- Building Enterprise AI: Hard-Won Lessons from 1200+ Hours of RAG Developmentbytevagabond.com
- What is Retrieval Augmented Generation (RAG)? | Databricksdatabricks.com
- Aman's AI Journal • Primers • Retrieval Augmented Generationaman.ai
- RAG Architecture Deep Divelinkedin.com
- Building A RAG Pipeline for Semi-structured Data with Langchain - Analytics Vidhyaanalyticsvidhya.com
- Code a simple RAG from scratchhuggingface.co