用词法检索搭配强大LLM，也能实现高效深度研究。
Rethinking Agentic Search with Pi-Serini: Is Lexical Retrieval Sufficient?
- 用BM25词法检索配合前沿LLM，构建搜索代理Pi-Serini。
- 在BrowseComp-Plus上达83.1%答案准确率，94.7%证据召回率。
- 调优和加深检索可显著提升效果，适合研究型系统设计者。
随着大语言模型在智能体循环中能力增强，词法检索是否仍足够？我们通过将BM25与具备更强推理和工具使用能力的前沿LLM结合，重新探讨该问题。为支持类似研究，提出Pi-Serini：一个配备检索、浏览、阅读三工具的搜索代理。实验显示，在BrowseComp-Plus数据集上，配置良好的词法检索器结合更强大的LLM（如gpt-5.5）即可实现有效深度研究，达到83.1%答案准确率和94.7%表面证据召回率，优于使用密集检索器的公开搜索代理。受控消融实验表明，优化BM25可使答案准确率提升18.0%，证据召回率提升11.1%；增加检索深度可使证据召回率再提升25.3%。源代码已开源。
原文摘要 · Abstract (English)
Does a lexical retriever suffice as large language models (LLMs) become more capable in an agentic loop? This question naturally arises when building deep research systems. We revisit it by pairing BM25 with frontier LLMs that have better reasoning and tool-use abilities. To support researchers asking the same question, we introduce Pi-Serini, a search agent equipped with three tools for retrieving, browsing, and reading documents. Our results show that, on BrowseComp-Plus, a well-configured lexical retriever with sufficient retrieval depth can support effective deep research when paired with more capable LLMs. Specifically, Pi-Serini with gpt-5.5 achieves 83.1% answer accuracy and 94.7% surfaced evidence recall, outperforming released search agents that use dense retrievers. Controlled ablations further show that BM25 tuning improves answer accuracy by 18.0% and surfaced evidence recall by 11.1% over the default BM25 setting, while increasing retrieval depth further improves surfaced evidence recall by 25.3% over the shallow-retrieval setting. Source code is available at https://github.com/justram/pi-serini.
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。