Oct 31, 08:39 PM

Patronus AI Launches Self-Serve API to Stop AI Hallucinations, 20% More Accurate with Reliability Guardrails

Patronus AI has launched the world's first self-serve API designed to stop AI hallucinations. The Patronus evaluation API offers powerful AI evaluation models, which are 20% more accurate than ragas in detecting hallucinations. Additionally, it surpasses Perspective and Llama Guard by 28% and 11% respectively in safety tasks. The API is particularly effective in practical domains such as finance and provides businesses with real-time guardrails and customizable evaluators to prevent AI failures. It also aims to accelerate AI development by equipping AI workloads with reliability guardrails.

#Patronus #Perspective #Llama Guard

Written with ChatGPT (GPT-4o).

Sources

Yangqing Jia@jiayq
2 years ago
Congrats to @PatronusAI @rebeccatqian et al for the launch! Custom models are not only more cost effective, but *better* than general purpose models. Can't wait to see awesome companies building vertical AI models and creating a vivid ecosystem. https://t.co/QbIC2gmbnt
AIwire@AIwireNews
2 years ago
Patronus AI Launches Self-Serve API for AI Evaluation and Guardrails https://t.co/g68ZeXAGxy @PatronusAI #AIwire #TCIwire
BigDATAWire@BigDATAwireNews
2 years ago
Patronus AI Launches Self-Serve API for AI Evaluation and Guardrails https://t.co/xl5pqN97fc @PatronusAI #BigDATAwire #datanami #TCIwire

Patronus AI Launches Self-Serve API to Stop AI Hallucinations, 20% More Accurate with Reliability Guardrails

Sources

Additional media

Similar Stories