ClusterBasedExampleSelector<T>

Models & Types ยท AiDotNet.PromptEngineering.FewShot

Selects examples using a clustering approach to ensure broad coverage.

Inheritance: FewShotExampleSelectorBase<T> โ†’ ClusterBasedExampleSelector<T>

For Beginners

Groups similar examples and picks one from each group.

Think of organizing photos:

  • Group 1: Beach photos
  • Group 2: Mountain photos
  • Group 3: City photos

Instead of showing all beach photos, you show one from each group.

Example:

Use this when:

  • Examples naturally fall into categories
  • You want guaranteed coverage of all categories
  • Building a general-purpose system

How It Works

This selector groups similar examples into clusters and selects representative examples from each cluster. This ensures broad coverage across different types of examples.

Constructors

Constructor Summary
ClusterBasedExampleSelector(Func<string, Vector<T>> embeddingFunction, int clusterCount, int? seed) Initializes a new instance of the ClusterBasedExampleSelector class.
ClusterBasedExampleSelector(IEmbeddingModel<T> embeddingModel, int clusterCount, int? seed) Initializes a new instance of the ClusterBasedExampleSelector class.

Properties

Property Summary
int ClusterCount { get; } Gets the number of clusters used for selection.