DCAgent3/medagentbench_rl__24GPU_base__exp_rpt_pymethods2test_large__GLM_4_7_swesmith_saf4a8f63a
收藏资源简介:
该数据集是一个包含对话记录的结构化数据集,主要用于人工智能代理(agent)和模型交互的跟踪与分析。数据集中的每个示例代表一次交互会话,包括对话内容(由角色和消息组成)、使用的代理、模型及其提供商、日期、任务类型、集数、运行ID、试验名称、结果、验证输出和跟踪来源。这些特征可用于研究多轮对话系统、模型性能评估或任务完成情况分析。数据集包含958个训练示例,总大小约36.6MB,适用于机器学习任务,如对话生成或代理行为模拟。
This dataset is a structured collection of conversation records, primarily designed for tracking and analyzing interactions between AI agents and models. Each example in the dataset represents an interaction session, including conversation content (composed of roles and messages), the agent used, model and its provider, date, task type, episode number, run ID, trial name, result, verifier output, and trace source. These features can be utilized for research on multi-turn dialogue systems, model performance evaluation, or task completion analysis. The dataset contains 958 training examples with a total size of approximately 36.6MB, suitable for machine learning tasks such as dialogue generation or agent behavior simulation.




