Skip to main content

​
Evaluator

A tool that measures the quality of materials generated by AI-powered educational applications. Evaluators assess specific aspects of content for pedagogical alignment and identify areas for improvement.

​
Rubric

A structured framework used to evaluate a concept based on learning science. It provides consistent criteria and serves as the foundation for a family of evaluators.
Illustration showing a sample rubric structure used by evaluators.

​
Dimension

A specific facet or attribute measured within a rubric. An individual evaluator typically focuses on a single rubric dimension.

​
Evaluator family

A collection of evaluators that score AI-generated content across multiple dimensions from one or more rubrics.

​
Accuracy

The degree to which an evaluator’s score aligns with curated or human-annotated validation datasets. Accuracy reflects the evaluator’s reliability and is expressed as a percentage.

​
Early access

An evaluator made available early in its development because it offers useful capabilities for research and experimentation. While stable, it remains limited in scope and under active development. Early-access evaluators invite feedback to guide iterative improvement.