PCToolkit: A Unified Plug-and-Play Prompt Compression Toolkit of Large Language Models

TL;DR

Evaluations of the compressors within PCToolkit across various natural language tasks, including reconstruction, summarization, mathematical problem-solving, question answering, few-shot learning, synthetic tasks, code completion, boolean expressions, multiple choice questions, and lies recognition are conducted.

Abstract

Prompt compression is an innovative method for efficiently condensing input prompts while preserving essential information. To facilitate quick-start services, user-friendly interfaces, and compatibility with common datasets and metrics, we present the Prompt Compression Toolkit (PCToolkit). This toolkit is a unified plug-and-play solution for compressing prompts in Large Language Models (LLMs), featuring cutting-edge prompt compressors, diverse datasets, and metrics for comprehensive performance evaluation. PCToolkit boasts a modular design, allowing for easy integration of new datasets and metrics through portable and user-friendly interfaces. In this paper, we outline the key components and functionalities of PCToolkit. We conducted evaluations of the compressors within PCToolkit across various natural language tasks, including reconstruction, summarization, mathematical problem-solving, question answering, few-shot learning, synthetic tasks, code completion, boolean expressions, multiple choice questions, and lies recognition.

Figures (2)

• Figure 1: Architecture of PCToolkit. The compressors module encompasses prompt compression methods that can be accessed through a unified interface with customizable parameters. The datasets module includes 10 diverse datasets detailed in Table \ref{['datasets']}. The metrics module comprises four primary metrics utilized for evaluating the performance of various compressors. The runner module offers a generalized interface for executing evaluations or simply retrieving the compressed prompt generated by the compressors.
• Figure 2: Demonstration website.