Evaluation

What is an Evaluation?

An Evaluation is essentially a test of the success or failure of an LLM call. Writing tests for Opper Functions is a great way to gain confidence in the quality of your LLM features.

Built in evaluation

Opper performs an automatic evaluation of each call through the platform. This is highlighted in the UI as a Score. Additionally each generation has an Observation tied to it where the reasoning of the evaluation is explained.