LLMSummarizationCompressor

Models & Types · AiDotNet.PromptEngineering.Compression

Compressor that uses an LLM to intelligently summarize and compress prompts.

Inheritance: PromptCompressorBase → LLMSummarizationCompressor

For Beginners

Uses AI to make your prompt shorter while keeping the meaning.

Example:

When to use:

  • For complex prompts where simple pattern matching won’t work
  • When semantic understanding is required
  • When maximum compression with preserved meaning is needed

How It Works

This compressor delegates compression to a language model, which can understand context and produce semantically equivalent but shorter versions of prompts. It’s the most intelligent form of compression but requires an LLM call.

Constructors

Constructor Summary
LLMSummarizationCompressor(Func<string, CancellationToken, Task<string>> summarizeFunc, string systemPrompt, Func<string, int> tokenCounter) Initializes a new instance of the LLMSummarizationCompressor class.

Properties

Property Summary
string SystemPrompt { get; } Gets the compression prompt template.

Methods

Method Summary
Task<string> CompressAsync(string prompt, CompressionOptions options, CancellationToken cancellationToken) Compresses the prompt asynchronously using the LLM.