Skip to content

Topic

Evaluation harnesses

Purpose-built software that runs a system against a set of tasks and scores the output, commonly used to measure models, agents and test suites.

Current clusters