TypeScript evaluation library for LLM applications. This package is vendor agnostic and can be used independently of any framework or platform.
Note
Note: This package is in alpha and subject to change.
Installation#
Note
@arizeai/phoenix-evals ships local docs and source code inside node_modules/@arizeai/phoenix-evals/.
Browse the website version under @arizeai/phoenix-evals Docs.
Start Here#
- Package overview for installation requirements, package layout, and module map
- Create evaluator for custom and code-based evaluators
- LLM evaluators for built-in model-backed evaluation
- Classification metrics for precision/recall/F-beta code evaluators
- Classification
- Templates
- Phoenix integration
Local Install Pattern#
After install, a coding agent can inspect:
That gives the agent version-matched prose plus the exact implementation and types that shipped with your project.