LLM-based agents

基于大语言模型的智能体，指由 LLM 驱动的、能够进行自然语言对话、推理和目标导向行为的自主或半自主软件系统，是当前人机交互研究的核心对象。

English Definition

"Autonomous or semi-autonomous software agents powered by large language models, capable of natural language dialogue, reasoning, and goal-directed behavior; a central object of study in Human-AI Interaction."

用法说明 · 针对中文母语者

HCI 中常和 rule-based agents、RL agents 对比，强调自然语言与推理能力。

常见错误

·中文常误译为「LLM 经纪人」——agent 在此指智能体而非经济角色
·和 chatbot 混用——agent 强调 goal-directed autonomy，chatbot 含义更泛

真实用例 · 2 条

"The landscape of human-AI negotiation has evolved from early rule-based systems to LLM-based agents capable of natural language dialogue and strategic reasoning"

论文Pipeline extracted — review needed·在综述中描述该领域的研究范式演进

"LLM-based agents add further variability: their behavior shifts under dimensional load"

论文Pipeline extracted — review needed·指出 LLM-based agents 引入的额外不确定性

由 pipeline 自动采集，待人工 review