Can Agents Generalize to the Open World? Unveiling the Fragility of Static Training in Tool Use

摘要

While Large Language Model (LLM) agents demonstrate proficiency in static benchmarks, their deployment in real-world scenarios is hindered by the dynamic nature of user queries, tool sets, and interaction dynamics. To address this generalization gap, we formalize **OpenAgent** (Tool-Use Agent in Open-World), a problem setting characterized by distributional shifts across query, action, observation, and domain dimensions. We construct a controlled sandbox environment where we define fine-grained environmental shifts across a four-tier hierarchy: *Perception*, *Interaction*, *Reasoning*, and *Internalization*. Our exhaustive analysis yields a series of key insights, demonstrating that agents trained via both Supervised Fine-Tuning and Reinforcement Learning suffer from varying degrees of performance degradation when confronting open environmental shifts. Building on these insights, we propose **Perturbation-Augmented Fine-Tuning**, a disturbance-based intervention strategy for SFT that lays the foundation for enhancing agent robustness and utility in realistic environments.