NVIDIA NeMo Guardrails integrates AutoAlign Sidecar to Protect LLMs at Scale with Sidecar Security
AutoAlign's novel Sidecar LLM firewall now secures custom generative AI chatbots via NVIDIA NeMo Guardrails, proving safety and security efficacy across all major LLMs — the details of which are described in a new white paper
Toronto, June 20, 2024 -- AutoAlign, an AI safety company that provides robust generative AI security, today announced that its dynamic Sidecar firewall is now available to secure custom chatbot AI models through NVIDIA NeMo Guardrails, part of the NVIDIA NeMo platform for building, customizing, and deploying generative AI models anywhere. The company also released a white paper highlighting its security and safety success on all major LLMs.
Sidecar technology runs alongside LLMs and provides consistent security by dynamically interacting with, while continually evolving alongside, rapidly growing models. With AutoAlign technology running on NVIDIA NeMo Guardrails, users can select Sidecar to secure custom, foundational chatbot models — allowing enterprises to confidently deploy powerful and secure generative AI.
"Legacy model security and safety solutions are failing to keep up with today’s risks, creating headaches for companies and major issues for customers,” said AutoAlign interim CEO and co-founder Dan Adamson. “Our white paper showcases that AI hazard mitigation should not be left to model providers, and Sidecar’s integration into NeMo Guardrails broadens the reach of our solution's impact.”
The company’s white paper exhibits that Sidecar’s guardrail architecture enhanced security on models like GPT-4 and Claude 3 Haiku. By deploying Sidecar with highly focused Alignment Controls, GPT-4 blocks the Garak LLM vulnerability scanner’s jailbreak attempts 100% of the time, up from 88.8%, and increases prompt injection mitigation from 14.3% to 100% security. Similarly, Sidecar improves Claude's jailbreak prevention from a 98.3% average to 100%, and prompt injection handling dramatically increases from 38.4% to 100%. NVIDIA NeMo Guardrails support for AutoAlign Sidecar technology helps ensure that LLMs leveraged to build custom chatbot applications meet rigorous enterprise demands while remaining secure and powerful.
"Sidecar, with highly tailored Alignment Controls, proves that it increases LLM security against widespread attacks, while also proactively preventing future issues,” said AutoAlign CTO and co-founder Rahm Hafiz. “Our white paper for system integration with NVIDIA NeMo Guardrails displays how Sidecar continually secures exponentially more powerful models while not degrading performance."
"Deploying generative AI models into chatbot applications can be a powerful tool for enterprises across every industry, and models need to be secure to deploy with confidence," said Amanda Saunders, director of Enterprise Generative AI software, NVIDIA. “With AutoAlign’s Sidecar running on NVIDIA NeMo Guardrails, developers can build and run generative AI models with enhanced protection.”
About AutoAlign
AutoAlign is an AI safety company that allows businesses to confidently deploy LLM and generative AI solutions. The company created Sidecar, the first fully automated guard rail firewall system that dynamically interacts with, and adapts alongside, LLMs hyper-evolution. This provides the most robust generative AI security that works comprehensively with every major model, so enterprises can leverage safe and effective AI solutions across their businesses. If your enterprise is ready to launch AI with confidence, reach out today: www.autoalign.ai
Contacts:
Peter Weltman
peter@mow.media
(415) 340-2040