ScholarPulse 日报 2026-09-29

2026-09-29 学术简报：2 篇。基于理论参数化的LLM生成式智能体能准确模拟并复现儿童的不同攻击行为亚型。

今日速览

序号	标题	来源	日期	主题	推荐等级
1	”Hot-Blooded” vs “Cold-Blooded”: Simulating the Behavioral Phenotypes of Childhood Aggression via Generative Agents	arXiv	2026-10-08	AI-Agent	高
2	Forms of LLM-Integrated Applications from LLM-Chats to Autonomous AI Agent System	arXiv	2026-10-08	RAG	高

重点论文与技术动态

1. “Hot-Blooded” vs “Cold-Blooded”: Simulating the Behavioral Phenotypes of Childhood Aggression via Generative Agents

来源：arXiv
日期：2026-10-08
作者/机构：Liping Fu
主题标签：AI-Agent, arXiv
推荐等级：高
分类：cs.HC

一句话结论

基于理论参数化的LLM生成式智能体能准确模拟并复现儿童的不同攻击行为亚型。

核心内容

检验了LLM智能体模拟儿童反应性、主动性及共现性攻击的建构效度。
相较非攻击基线具强区分效度和大效应量，高跨种子信度证实行为分化源于潜在心理参数而非模型随机性。
定性叙事分析与现有实证文献相吻合，证实智能体可准确复现不同攻击亚型。

方法与数据

基于社会信息加工模型参数化4个智能体，在8个社交场景中运行1,920次模拟，并采用混合盲编码流水线提取32个定量行为指标。

价值判断

值得关注：为假设生成、干预试点和心理测量工具优化提供了可扩展、高可控的框架。
可复用点：理论驱动的智能体参数化方法与提取行为指标的混合盲编码流水线。
局限/待核查：摘要未明确具体使用的底层LLM与社交场景细节。

摘要

This study examines the construct validity of LLM-based generative agents in simulating reactive, proactive, and co-occurring aggression in children. Four distinct agents were instantiated using a theory-driven parameterization grounded in the social information processing model. A total of 1,920 simulation runs were conducted across eight social scenarios, employing a hybrid blind-coding pipeline to extract 32 quantitative behavioral indicators. Results demonstrate robust discriminant validity relative to a non-aggressive baseline, with large effect sizes. High cross-seed reliability confirms that behavioral differentiation is driven by underlying psychological parameters rather than model stochasticity. Qualitative narrative analyses further converged with established empirical literature. Overall, these findings indicate that theory-parameterized LLM agents can accurately reproduce distinct aggression subtypes, offering a scalable, highly controllable framework for hypothesis generation, intervention piloting, and the refinement of psychological measurement tools.

2. Forms of LLM-Integrated Applications from LLM-Chats to Autonomous AI Agent System

来源：arXiv
日期：2026-10-08
作者/机构：Irene Weber
主题标签：RAG, arXiv
推荐等级：高
分类：cs.CL, cs.AI, cs.SE

一句话结论

本综述系统评估了LLM集成应用的各类标签，证实其具备真实的软件架构内涵，并归纳了七种常见应用形态。

核心内容

厂商标签含架构内容：Copilot指需用户逐步确认以操作宿主应用的路由-工作器架构，Agent则对应AI规划多步执行且用户仅见结果的架构。
四大主流供应商的编程智能体共享“推理-行动”循环并委派给子智能体的单一架构。
以智能体与工具为通用词汇，定义了LLM聊天、自定义智能体、RAG、AI增强工作流、Copilot、编程智能体及智能体RAG七种形态。

方法与数据

沿架构模式、执行控制与用户干预点、单任务智能体调用次数、工具使用四个结构维度进行特征刻画。
采用源自研究出版物与厂商文档的22个系统作为说明性语料库。

价值判断

值得关注：厘清了行业营销标签背后的真实系统架构差异。
可复用点：用于界定LLM应用的四个结构维度及通用词汇框架。
局限/待核查：智能体RAG仅被部分刻画；22个系统的语料库也展示了上述分类描述达到极限的边界。

摘要

Large language models (LLMs) are increasingly embedded as components in software systems, marketed under labels such as chatbot, copilot, retrieval-augmented generation, workflow, coding agent and AI agent. Whether these labels denote genuine architectural forms or serve as branding has not been assessed systematically. In the sources surveyed, labels do carry architectural content, most clearly in vendor usage: copilot denotes a router-worker architecture operating a host application under step-by-step user confirmation, while the more recent shift to the label agent coincides with AI-planned multi-step execution of which the user sees only the outcome. The coding agents of four major providers share one architecture, a reason-and-act loop delegating to subagents. This survey describes seven recurring forms---LLM chats, custom agents, retrieval-augmented generation (RAG), AI-enhanced workflows, copilots, coding agents, and, in part, agentic RAG---in a common vocabulary of agents and tools. Each is characterized along four structural dimensions (agentic RAG only partially): the architectural pattern, the control of execution and the point of user intervention, the number of agent calls per task, and tool use. An illustrative corpus of 22 systems from research publications and vendor documentation grounds the descriptions and shows where they reach their limit.