Learning Simple Test-Time Environments for LLM Web Agents

2026-08-29

AI Summary

为解决LLM代理在复杂现实环境中的性能下降问题，提出一种无标签学习方法TTED，在测试时分解环境观察以提升代理行为适应性。

Abstract

Large language model (LLM) agents have demonstrated remarkable proficiency in manually constructed environments, yet their performance frequently collapses when transitioned to complex real-world settings. Existing research largely attribute this degradation to the compositional generalization gaps in LLMs on combinations of multiple simple, well-structured environments. In this work, we propose that LLM web agents can learn simple environment observations at test time. Specifically, we introduce trial steps for agents to decompose a complex environment observation into sub-modules, and implement a label-free learning method, Test-Time Environment Decomposition (TTED), to adapt agent behaviors with experience during inference. Our empirical evaluations demonstrate the framework's efficacy across both synthetic and realistic benchmarks, showing (1) experience gains acquired within simpler sub-environments can be effectively composed to improve performance in the full one, and (2) test-time training on sub-environments can significantly enhance the compositional generalization of agents in real-world web automation tasks. We also provide key insights in the design of the label-free learning algorithm. As more complex environments are accessed by LLM agents, we believe learning environment decomposition skills at test time will be critical for robust real-world deployment.

Problem

• Large language model (LLM) agents
• performance degradation
• compositional generalization gaps
• real-world settings

Innovation

• Test-Time Environment Decomposition
• label-free learning
• compositional generalization
• web automation