NeMo Guardrails

Guardrails AI adds programmable guardrails to large language models, ensuring reliability and safety through input/output validation, structured data extraction, and custom validators.

CVS Health's open-source uncertainty quantification library for language models, providing UQ-based hallucination detection with confidence scoring and mitigation tools to identify and reduce unreliable LLM outputs.

OpenAI's framework for evaluating LLMs and LLM systems, providing an open-source registry of benchmarks and tools for systematic model assessment.

An open-source evaluation and testing library for LLM agents providing automated model scanning, bias detection, performance benchmarking, and compliance checks.