The word 'AutoRail' is autoGen and guardrail combined. Make LLM's output more consistent with human values, using AutoGen and Guardrails.

AutoRail is a proof of concept that utilizes AutoGen and Guardrails to align the output of LLMs with human values. It is a project that is still in progress. Our goal is to build a complete AI safety framework to evaluate and modify the output text of LLM so that its values align with those of governments, political parties, and other organizations.

This project provides a framework to score, evaluatr, and revise the output of LLM. We provide a framework and you provide the value documents of your organization. This ¹

This example can be run with

python autorail/autorail.py

• Create a proof of concept
• AutoRail group chat
• Log messages and token cost
• Convert guardrails to system prompt
• Azure deployment
• Write framework details
• Write tutorials
• Package it as a library

Get started by checking out Github issues and of course using Guardrails to familiarize yourself with the project. We encourage users to provide public value documents. Convert your value documents into a format that can be used by guardrails.