LLM Agents Factory：领域特定 LLM 代理的检索框架
原标题：LLM Agents Factory: Retrieval of Domain-Specific LLM Agents
AI 摘要
LLM Agents Factory 是一个基于检索的框架，利用超过 20K 个预定义代理配置文件，按需构建领域特定且基于 Wikipedia 的代理。该框架支持语义搜索检索和蒸馏两种模式，在 MMLU、BIG-bench 等基准测试中，其检索式代理构建在准确率上超越非代理基线，并以更低推理成本匹配 AutoGen 的生成质量。该研究将代理构建视为信息检索问题，为工业应用提供了成本高效、准确且可控的替代方案。论文将发表于 ACM SIGIR 2026。
正文节选
LLM Agents Factory: Retrieval of Domain-Specific LLM Agents Abstract. Large language model (LLM) agents improve task performance by decomposing problems into role-specialized behaviors. However, their practical deployment is often limited by the computational cost and instability associated with the on-the-fly agent design for each user request. To address this, we present LLM Agents Factory, a retrieval-based framework that constructs domain-specific and Wikipedia-grounded agents on demand usin