Скачать или смотреть Basics of LLM Testing - DeepEval

In this video, we take a quick look at DeepEval, a powerful tool for evaluating LLM outputs and RAG (Retrieval-Augmented Generation) contexts. We’ll explore some of the basic metrics it provides to help assess the quality and relevance of your results.

⚠️ Important Note: Always use a different LLM for evaluation than the one used in your main workflow or agent to ensure unbiased testing. Threshold allows you to control the acceptance of a test.


0:00 - Intro
1:29 - GEval
1:54 - Contextual, Faithfulness, Contextual Metric
2:21 - DAG
3:15 - GEval Example
5:15 - Faithfulness Example
7:28 - Contextual Precision Example
9:21 - Answer Relevancy Example