智能体交互对话数据集

本数据集聚焦AI智能体之间的自然语言交互对话，覆盖任务协作、信息共享、协商谈判、社交闲聊、辩论、教学辅导等场景。数据内容包括会话元数据（主题、场景、参与角色、任务完成状态）、对话轮次（发言者、发言内容、时间戳）、行为意图标签（提问/陈述/确认/建议/拒绝等）、对话质量评估（相关性、信息增益、流畅度）及状态追踪。适用于多智能体对话策略学习、对话一致性评测、协作效率分析、对话生成模型微调及社交智能体行为建模等场景。This dataset centers on natural language interactive conversations among AI Agents, covering scenarios including task collaboration, information sharing, negotiation, social small talk, debate, and instructional tutoring. The dataset content includes conversation metadata (topic, scenario, participating roles, task completion status), conversation turns (speaker, utterance content, timestamp), behavioral intention tags (such as query, statement, confirmation, suggestion, rejection), conversation quality assessment (relevance, information gain, fluency), and state tracking. It is applicable to scenarios such as multi-agent conversation strategy learning, conversation consistency evaluation, collaboration efficiency analysis, fine-tuning of conversation generation models, and behavior modeling of social AI Agents.

搜集汇总

背景与挑战

背景概述

该数据集专注于AI智能体之间的自然语言交互对话，覆盖任务协作、信息共享、协商谈判、社交闲聊、辩论及教学辅导等多种场景。数据包含会话元数据、对话轮次、行为意图标签和质量评估等结构化内容，适用于多智能体对话策略学习、对话一致性评测以及对话生成模型微调等应用。