What is AutoRAG?
AutoRAG is an advanced RAG platform for knowledge driven chatbots.
It overcomes various difficult failure modes across content ingestion, chunking, metadata & complex docs/queries not handled by most RAG frameworks. It is an opinionated RAG pipeline with highly optimized specialized task models that serve small context length LMs as well as long context length strong LLMs.
AutoRAG comes with TruthChecker hallucination detection to minimize LLM hallucinations.
Features
-
Knowledge Base connectors to custom data sources and documents (structured and semi structured including PDFs, Slide decks and Web pages
-
Choice of LLMs: FLAN UL-2, Llama 2, MosaicML, ChatGPT, GPT-4
-
TruthChecker AI fine tuned for detecting hallucinations in multi-turn responses
-
Product UI or Microservice API for customer specific UIs
-
Fluid conversational dialog with hallucination/fact checking and disambiguation
-
Active learning, fine tuning as the solution continues to run to achieve highest accuracy
-
On-prem deployment with Open Source LLMs
-
Integrations into Agent Assist platforms
3 steps to a transformative solution for Copilots and CX Autopilots
Natural Language Knowledge Retrieval
Bring your internal documents and knowledge bases and get an Agent Copilot bot that your agents can use to optimize their performance.
Real Time Fact Checking
Agent Copilot uses Got It AI's TruthChecker technology to reduce LLM hallucinations to low, GPT-4 levels
Human Level Accuracy
AutoRAG comes with a sophisticated active learning system that continuously improves performance, reaching human level or better accuracy