Abstract
Large Language Model (LLM) agents show great promise for complex, multi-turn tool-use tasks, but their development is often hampered by the extreme scarcity of high-quality training data. Supervised fine-tuning (SFT) on synthetic data leads to overfitting, whereas standard reinforcement learning (RL) struggles with a critical cold-start problem and training instability. To address these challenges, we introduce $\textbf{Environment Tuning}$, a novel training paradigm that enables agents to learn complex behaviors directly from problem instances without relying on pre-collected expert trajectories. $\textbf{Environment Tuning}$ orchestrates this learning process through a structured curriculum, actionable environment augmentation that provides corrective feedback, and fine-grained progress rewards to ensure stable and efficient exploration. Using only 400 problem instances from Berkeley Function-Calling Leaderboard (BFCL) benchmark, our method not only achieves competitive in-distribution performance against strong baselines but also demonstrates superior out-of-distribution generalization, overcoming the performance collapse common to SFT-based approaches. Our work presents a paradigm shift from supervised fine-tuning on static trajectories to dynamic, environment-based exploration, paving the way for training more robust and data-efficient agents.
Abstract (translated)
大型语言模型（LLM）代理在复杂的多轮工具使用任务中展现出巨大潜力，但其开发常常受限于高质量训练数据的极度稀缺。基于合成数据的监督微调（SFT）会导致过拟合，而标准强化学习（RL）则面临严重的冷启动问题和训练不稳定的问题。为解决这些问题，我们引入了**环境调优**，这是一种新的培训范式，它使代理能够直接从问题实例中学习复杂行为，而不依赖于预先收集的专家轨迹。**环境调优**通过一个结构化的课程、可操作的环境增强（提供纠正反馈）和细致的进步奖励来组织这个学习过程，以确保稳定而高效的探索。 仅使用伯克利函数调用排行榜（BFCL）基准测试中的400个问题实例，我们的方法不仅在分布内性能上与强大的基线模型相匹敌，并且展示了超出分布的优越泛化能力，克服了基于SFT的方法常见的性能崩溃问题。我们的工作标志着从静态轨迹上的监督微调向动态、环境导向探索的根本转变，为训练更稳健和数据高效代理铺平道路。
URL
https://arxiv.org/abs/2510.10197
