WebLeaper: Empowering Efficiency and Efficacy in WebAgent via Enabling Info-Rich Seeking

TL;DR

WebLeaper introduces a novel framework for web information seeking that enhances search efficiency and effectiveness by formulating IS as a tree-structured reasoning problem and synthesizing high-coverage tasks using curated data, leading to improved performance on multiple benchmarks.

Contribution

The paper presents WebLeaper, a new approach that constructs high-coverage IS tasks and models search as tree-structured reasoning to improve efficiency and efficacy of LLM-based agents.

Findings

Significant improvements in search efficiency and effectiveness on five benchmarks.

WebLeaper's task synthesis methods outperform baseline approaches.

Enhanced generalization of search behaviors in LLM agents.

Abstract

Large Language Model (LLM)-based agents have emerged as a transformative approach for open-ended problem solving, with information seeking (IS) being a core capability that enables autonomous reasoning and decision-making. While prior research has largely focused on improving retrieval depth, we observe that current IS agents often suffer from low search efficiency, which in turn constrains overall performance. A key factor underlying this inefficiency is the sparsity of target entities in training tasks, which limits opportunities for agents to learn and generalize efficient search behaviors. To address these challenges, we propose WebLeaper, a framework for constructing high-coverage IS tasks and generating efficient solution trajectories. We formulate IS as a tree-structured reasoning problem, enabling a substantially larger set of target entities to be embedded within a constrained context. Leveraging curated Wikipedia tables, we propose three variants for synthesizing IS tasks, Basic, Union, and Reverse-Union, to systematically increase both IS efficiency and efficacy. Finally, we curate training trajectories by retaining only those that are simultaneously accurate and efficient, ensuring that the model is optimized for both correctness and search performance. Extensive experiments on both basic and comprehensive settings, conducted on five IS benchmarks, BrowserComp, GAIA, xbench-DeepSearch, WideSearch, and Seal-0, demonstrate that our method consistently achieves improvements in both effectiveness and efficiency over strong baselines.