让小模型学会可靠搜索，比大模型更准更省
Search, Do not Guess: Teaching Small Language Models to Be Effective Search Agents
- 轻量微调让小模型学会依赖检索结果，不靠猜
- 在Bamboogle和HotpotQA上分别提升17.3和15.3分
- 适合想用小模型做精准问答的开发者
具备搜索工具的智能体已成为解决知识密集型任务的有效方案。尽管大语言模型（LLMs）具备强大的推理能力，但其高计算成本限制了其在搜索代理中的实际部署。因此，近期研究聚焦于将大模型的智能体行为蒸馏到小语言模型（SLMs）中。通过在复杂多跳推理任务上的全面评估，我们发现尽管参数量较少，但SLMs调用搜索工具的频率更低，且更容易产生幻觉。为解决此问题，我们提出 extit{policy}，一种轻量级微调方法，明确训练SLMs依赖检索证据进行生成。相比从LLM蒸馏的代理，该方法在Bamboogle上提升17.3分，在HotpotQA上提升15.3分，实现了与大模型相当的性能表现。进一步分析表明，SLMs中自适应搜索策略常导致性能下降，凸显出一致搜索行为对可靠推理的重要性。
原文摘要 · Abstract (English)
Agents equipped with search tools have emerged as effective solutions for knowledge-intensive tasks. While Large Language Models (LLMs) exhibit strong reasoning capabilities, their high computational cost limits practical deployment for search agents. Consequently, recent work has focused on distilling agentic behaviors from LLMs into Small Language Models (SLMs). Through comprehensive evaluation on complex multi-hop reasoning tasks, we find that despite possessing less parametric knowledge, SLMs invoke search tools less frequently and are more prone to hallucinations. To address this issue, we propose \policy, a lightweight fine-tuning approach that explicitly trains SLMs to reliably retrieve and generate answers grounded in retrieved evidence. Compared to agent distillation from LLMs, our approach improves performance by 17.3 scores on Bamboogle and 15.3 scores on HotpotQA, achieving LLM-level results across benchmarks. Our further analysis reveals that adaptive search strategies in SLMs often degrade performance, highlighting the necessity of consistent search behavior for reliable reasoning.
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。