Models model language model announce available analysis trend

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

arXiv cs.CLby [Submitted on 2 Apr 2026]April 4, 20262 min read1 views

arXiv:2604.01535v1 Announce Type: new 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

View PDF HTML (experimental)

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.

Subjects:

Computation and Language (cs.CL)

Cite as: arXiv:2604.01535 [cs.CL]

(or arXiv:2604.01535v1 [cs.CL] for this version)

https://doi.org/10.48550/arXiv.2604.01535

arXiv-issued DOI via DataCite (pending registration)