Evaluate
Overview
Stop experimenting in spreadsheets and notebooks. Use Evaluate’s powerful insights to build GenAI systems that just work.
Galileo Evaluate is a powerful bench for rapid, collaborative experimentation and evaluation of your LLM applications.
Core features
-
Tracing and Visualizations - Track the end-to-end execution of your queries. See what happened along the way and where things went wrong.
-
State-of-the-art Metrics - Combine our research-backed Guardrail Metrics with your own Custom Metrics to evaluate your system.
-
Experiment Management - Track all your experiments in one place. Find the best configuration for your system.
An Evaluation Run of a RAG Workflow
The Workflow
1
Log your runs
Integrate promptquality into your system or test a template model combination through the Playground. Choose and register your metrics to define what success means for your use case.
2
Analyze results
Identify poor perfomance, trace it to the broken step, form hypothesis on what could be behind it.
3
Debug, Fix & Run another Eval
Tweak your system and try again until your quality bar is met.
Getting Started
Was this page helpful?