Read More, Think More: Revisiting Observation Reduction for Web Agents

TL;DR

This paper investigates how the choice of web page observation representations affects large language model-based web agents, emphasizing the importance of adapting observations to model capability and token budget.

Contribution

It provides empirical insights and practical guidelines for selecting observation representations based on model capability and token constraints.

Findings

Higher-capability models benefit from detailed HTML observations.

Lower-capability models perform better with compact observations like accessibility trees.

Observation history and diff-based representations improve performance.

Abstract

Web agents based on large language models (LLMs) rely on observations of web pages -- commonly represented as HTML -- as the basis for identifying available actions and planning subsequent steps. Prior work has treated the verbosity of HTML as an obstacle to performance and adopted observation reduction as a standard practice. We revisit this trend and demonstrate that the optimal observation representation depends on model capability and thinking token budget: (1) compact observations (accessibility trees) are preferable for lower-capability models, while detailed observations (HTML) are advantageous for higher-capability models; moreover, increasing thinking tokens further amplifies the benefit of HTML. (2) Our error analysis suggests that higher-capability models exploit layout information in HTML for better action grounding, while lower-capability models suffer from increased hallucination under longer inputs. We also find that incorporating observation history improves performance across most models and settings, and a diff-based representation offers a token-efficient alternative. Based on these findings, we suggest practical guidelines: adaptively select observation representations based on model capability and thinking token budget, and incorporate observation history using diff-based representations.