用信息瓶颈方法优化网页智能体，大幅降本仅小幅损精度。
WebRouter: Query-specific Router via Variational Information Bottleneck for Cost-sensitive Web Agent
- 基于信息瓶颈设计查询专用路由，压缩输入并控制成本。
- 在5个真实网站上降低87.8%操作成本，准确率仅降3.8%。
- 适合追求高性价比的网页自动化应用开发者。
LLM驱动的网页智能体虽具备强大自动化能力，但面临成本与性能的权衡难题。其复杂提示（包含目标、动作历史和环境状态）导致大模型集成性能下降。为此，我们提出WebRouter，一种从信息论视角训练的查询特定路由机制。核心是引入成本感知的变分信息瓶颈（ca-VIB）目标，在学习输入提示压缩表示的同时，显式惩罚预期操作成本。在WebVoyager基准的五个真实网站上的实验表明，相比GPT-4o基线，WebRouter将操作成本降低87.8%，准确率仅下降3.8%。
原文摘要 · Abstract (English)
LLM-brained web agents offer powerful capabilities for web automation but face a critical cost-performance trade-off. The challenge is amplified by web agents' inherently complex prompts that include goals, action histories, and environmental states, leading to degraded LLM ensemble performance. To address this, we introduce WebRouter, a novel query-specific router trained from an information-theoretic perspective. Our core contribution is a cost-aware Variational Information Bottleneck (ca-VIB) objective, which learns a compressed representation of the input prompt while explicitly penalizing the expected operational cost. Experiments on five real-world websites from the WebVoyager benchmark show that WebRouter reduces operational costs by a striking 87.8\% compared to a GPT-4o baseline, while incurring only a 3.8\% accuracy drop.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。