Discrete Prompt Compression with Reinforcement Learning

Abstract

Compressed prompts aid instruction-tuned language models (LMs) in overcoming context window limitations and reducing computational costs. Existing methods, which primarily based on training embeddings, face various challenges associated with interpretability, the fixed number of embedding tokens, reusability across different LMs, and inapplicability when interacting with black-box APIs. This study proposes prompt compression with reinforcement learning (PCRL), which is a discrete prompt compression method that addresses these issues. The proposed PCRL method utilizes a computationally efficient policy network that edits prompts directly. The training approach employed in the proposed PCRLs can be applied flexibly to various types of LMs, including both decoder-only and encoder-decoder architecture and it can be trained without gradient access to the LMs or labeled data. The proposed PCRL achieves an average reduction of 24.6% in terms of the token count across various instruction prompts while maintaining sufficient performance. In addition, we demonstrate that the learned policy can be transferred to larger LMs, and through a comprehensive analysis, we explore the token importance within the prompts. Our code is accessible at https://github.com/nenomigami/PromptCompressor.

References in corpus (8)

• Llama 2: Open Foundation and Fine-Tuned Chat Models
• Scaling Instruction-Finetuned Language Models
• ChatGPT Outperforms Crowd-Workers for Text-Annotation Tasks
• Is ChatGPT better than Human Annotators? Potential and Limitations of ChatGPT in Explaining Implicit Hate Speech
• The Falcon Series of Open Language Models
• The Summary Loop: Learning to Write Abstractive Summaries Without Examples
• Learning to Compress Prompts with Gist Tokens
• Unlocking Context Constraints of LLMs: Enhancing Context Efficiency of LLMs with Self-Information-Based Content Filtering