Cover Image for Evaluate LLMs with Evidently open-source
Cover Image for Evaluate LLMs with Evidently open-source
Hosted By
63 Going

Evaluate LLMs with Evidently open-source

Hosted by Evidently AI
Registration Closed
This event is not currently taking registrations. You may contact the host or subscribe to receive updates.
About Event

Evidently is an open-source tool to evaluate, test, and monitor ML-powered systems.

Join our Evidently Community Call for an in-depth walkthrough of the latest open-source features. You will learn how to effectively evaluate, test, and monitor LLM-powered applications, including RAGs and chatbots.

Deep dive with a code demo. We will show a code walkthrough explaining all the steps needed to design and run evaluations for a sample LLM system. We will cover several evaluation methods, including:

  • LLM as a judge

  • Model-based evaluations

  • Semantic similarity 

  • Regular expressions 

  • Drift detection 

You will learn how to generate visual Reports summarizing evaluation results, run Test Suites for regression testing, and get a live Monitoring Dashboard to track evaluation results in time. 

We will explain how you can customize such evaluations to your use - to detect toxicity, bias, leakage of personal data, or other properties that define AI quality for your use case. We will also discuss the pros and cons of different methods and answer your questions.

Hosted By
63 Going