通过有效令牌成本分析深度搜索中的重排序权衡，发现适度重排序比增加推理更高效
Rerank Before You Reason: Analyzing Reranking Tradeoffs through Effective Token Cost in Deep Search Agents
- 提出有效令牌成本（ETC）度量，评估重排序与推理的资源分配
- 适度重排序可提升检索与端到端准确率，且成本远低于增加推理
- 适用于追求高效率的复杂问题求解系统设计者
深度研究代理依赖迭代检索与推理来回答复杂问题，但测试时计算量的扩展带来了显著的效率挑战。本文聚焦于深度搜索流程中推理预算的分配，重点研究列表级重排序的作用。基于BrowseComp-Plus基准，我们通过一种新颖的有效令牌成本（ETC）指标，分析模型规模、推理投入、重排序深度与总令牌成本之间的权衡。结果表明，重排序能持续提升检索与端到端准确率；适度重排序通常比增加搜索时推理带来更大收益，在实现相近准确率的同时显著降低整体成本。所有代码已开源。
原文摘要 · Abstract (English)
Deep research agents rely on iterative retrieval and reasoning to answer complex queries, but scaling test-time computation raises significant efficiency concerns. We study how to allocate reasoning budget in deep search pipelines, focusing on the role of listwise reranking. Using the BrowseComp-Plus benchmark, we analyze tradeoffs between model scale, reasoning effort, reranking depth, and total token cost via a novel effective token cost (ETC) metric. Our results show that reranking consistently improves retrieval and end-to-end accuracy, and that moderate reranking often yields larger gains than increasing search-time reasoning, achieving comparable accuracy at substantially lower cost. All our code is available at https://github.com/sahel-sh/DeepHone
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。