Skip to main content
These are several common ways that edtech teams use Evaluators to vet, monitor, and improve their generated AI content.

​
Vet content quality in production

Ensure that your AI-generated text passage properly caters to its intended grade level before shipping:
  1. Assess the grade level for your text passage
  2. Rewrite the passage if it doesn’t align with your target grade
  3. Track how often your model must rewrite its output to determine whether you should refine your prompts
Follow the Tutorial for step-by-step implementation.

​
Monitor quality in production

Monitor shifts in the quality of your models before they affect students at scale:
  1. Score your models’ outputs before and after any change
  2. Track scores over time to anticipate any issues
Follow the Tutorial for step-by-step implementation.

​
Improve quality by refining prompts

Make data-driven and measurable improvements to your prompts:
  1. Score your models’ outputs before and after a prompt change
  2. Use evaluators’ reasoning to make targeted improvements to prompts
Follow the Tutorial for step-by-step implementation.

Resource Library - Evaluators

Find guides and use cases for Evaluators.

Quickstart

Run an evaluator in the Evaluators playground, a Python notebook, or with the SDK.