AdaRefiner: Refining Decisions of Language Models with Adaptive Feedback

TL;DR

AdaRefiner is a framework that improves large language models' decision-making by automatically refining their understanding through reinforcement learning feedback, reducing the need for manual prompt engineering and fine-tuning.

Contribution

It introduces a lightweight adapter that enables automatic self-refinement of LLMs with RL feedback, enhancing decision-making in complex tasks without extensive prompt engineering.

Findings

Outperforms baselines on 22 open-world game tasks

Enhances high-level and common-sense skills in agents

Reduces need for prompt engineering and fine-tuning

Abstract

Large Language Models (LLMs) have demonstrated significant success across various domains. However, their application in complex decision-making tasks frequently necessitates intricate prompt engineering or fine-tuning, leading to challenges in unseen downstream tasks and heavy demands on computational resources. Meanwhile, Reinforcement Learning (RL) has been recognized as effective in decision-making problems but struggles in environments with sparse rewards, such as open-world games. To overcome these challenges, we introduce AdaRefiner, a novel framework designed to enhance the synergy between LLMs and RL feedback. The key component of AdaRefiner is a lightweight Adapter Language Model (LM), which automatically refines task comprehension based on feedback from RL agents. This method mitigates the need for intricate prompt engineering and intensive LLM fine-tuning while maintaining the LLMs' generalization abilities and enhancing their decision-making capabilities in downstream tasks. Empirical evaluations of AdaRefiner on 22 diverse tasks within the open-world game Crafter have demonstrated its superior effectiveness, especially in guiding agents towards higher-level and common-sense skills. Our work makes contributions to the automatic self-refinement of LLMs with RL feedback, offering a more adaptable and efficient solution for complex decision-making problems.