小模型通过动态构建任务向量实现高效网页搜索，性能媲美大模型。
Lucy: edgerunning agentic web search on mobile with machine generated task vectors
- 将思考过程视为动态任务向量生成机制，实时构造推理路径。
- 在SimpleQA上达78.3%准确率，与深求大模型相当。
- 适合移动端轻量化智能搜索场景，尤其适用于资源受限设备。
小型语言模型（SLMs）因容量限制，在知识密集型任务中表现受限。尽管测试时计算可提升性能，但多数方法将推理视为固定或启发式过程。本文提出新范式：将模型内部以标记的推理过程，视为动态任务向量机器。不将其仅看作思维痕迹，而是将其生成过程本身理解为模型即时构建与优化自身任务向量的机制。我们通过RLVR优化该动态任务向量机器，成功训练出一个代理型网页搜索模型——Lucy。该模型仅1.7B参数，结合MCP集成，可在SimpleQA基准上达到78.3%准确率，性能媲美深求-3等大型模型。这表明，当具备结构化自构建任务推理能力时，小模型亦可媲美大模型。
原文摘要 · Abstract (English)
Small language models (SLMs) are inherently limited in knowledge-intensive tasks due to their constrained capacity. While test-time computation offers a path to enhanced performance, most approaches treat reasoning as a fixed or heuristic process. In this work, we propose a new paradigm: viewing the model's internal reasoning, delimited by <think> and </think> tags, as a dynamic task vector machine. Rather than treating the content inside these tags as a mere trace of thought, we interpret the generation process itself as a mechanism through which the model \textbf{constructs and refines its own task vectors} on the fly. We developed a method to optimize this dynamic task vector machine through RLVR and successfully trained an agentic web-search model. We present Lucy, a 1.7B-parameter SLM that leverages this dynamic reasoning mechanism with MCP integration to achieve 78.3% accuracy on the SimpleQA benchmark, performing on par with much larger models such as DeepSeek-V3. This demonstrates that small models can rival large ones when equipped with structured, self-constructed task 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。