Arthur Shield
LLM firewall and evaluation from an ML observability vendor.
Made by ArthurAI guardrails
About Arthur Shield
The terms of service are a legal agreement between users and ArthurAI, Inc. Arthur helps enterprise teams optimize model operations and performance at scale. Arthur's discovery capabilities find every agent running on endpoints, cloud, private data centers, and anywhere else in your compute environment.
Arthur Observability can be used across LLM, tabular, CV, and NLP models to monitor, measure, and improve model metrics and achieve business results. Track spending by team and user. Findings can be streamed to the tools a security operations center already uses.
Drafted from arthur.ai · not yet confirmed by Arthur
How it works
Arthur Shield checks prompts and responses for hallucination, toxicity, PII, prompt injection and off policy content, and Arthur's wider platform adds evaluation, benchmarking and monitoring of models in production. Its background is model observability for banks and insurers, so it's strong on the audit and monitoring side. Parts of its evaluation engine are open source.
Reviews
Be the first to review it
AnonymousKnow a team weighing Arthur Shield? Send them this page.