Bootstrapping Exploration with Group-Level Natural Language Feedback in Reinforcement Learning
This work addresses the problem of inefficient exploration in reinforcement learning for large language models by better utilizing natural language feedback, which is an incremental improvement to existing RL algorithms.
This paper introduces GOLF, a reinforcement learning framework that leverages group-level natural language feedback to enhance exploration efficiency. By aggregating external critiques and intra-group attempts, GOLF generates high-quality refinements that are injected into training, achieving a 2.2x improvement in sample efficiency compared to RL methods using only scalar rewards.
Large language models (LLMs) typically receive diverse natural language (NL) feedback through interaction with the environment. However, current reinforcement learning (RL) algorithms rely solely on scalar rewards, leaving the rich information in NL feedback underutilized and leading to inefficient exploration. In this work, we propose GOLF, an RL framework that explicitly exploits group-level language feedback to guide targeted exploration through actionable refinements. GOLF aggregates two complementary feedback sources: (i) external critiques that pinpoint errors or propose targeted fixes, and (ii) intra-group attempts that supply alternative partial ideas and diverse failure patterns. These group-level feedbacks are aggregated to produce high-quality refinements, which are adaptively injected into training as off-policy scaffolds to provide targeted guidance in sparse-reward regions. Meanwhile, GOLF jointly optimizes generation and refinement within a unified RL loop, creating a virtuous cycle that continuously improves both capabilities. Experiments on both verifiable and non-verifiable benchmarks show that GOLF achieves superior performance and exploration efficiency, achieving 2.2$\times$ improvements in sample efficiency compared to RL methods trained solely on scalar rewards. Code is available at https://github.com/LuckyyySTA/GOLF.