Cách chạy đánh giá trong giao diện người dùng LangSmith

Roles like product managers or subject matter experts typically have the most context on the quality and performance of an LLM application, and thus can play a crucial role in leading the evaluation and improving AI applications.

In this video, we will show you how a LangSmith user, like PM or subject matter expert, can walk through the evaluation flow in the LangSmith UI.

0:00 - Introduction

0:39 - Steps to Running an Offline Evaluation

1:38 - Creating a Prompt

2:43 - Creating a Dataset

3:30 - Defining Evaluators

6:50 - Running & Visualizing an Experiment