DCAgent3/dev_set_v2_rl__24GPU_base_lr5e_6__exp_rpt_pymethods2test_large__GLM_4_7_swesmit39f22783
收藏数据链接:
官方服务:
资源简介:
该数据集包含多轮对话记录及相关元数据,用于记录和分析AI代理在特定任务中的交互表现。数据集包含以下特征:对话内容(角色和内容)、代理信息、模型名称及提供商、日期、任务类型、执行轮次、运行ID、试验名称、执行结果、验证器输出和追踪来源。数据集共有440个训练样本,适用于对话系统评估、代理行为分析和任务执行结果研究。
This dataset contains multi-turn conversation records and related metadata, used to record and analyze the interaction performance of AI agents in specific tasks. The dataset includes the following features: conversations (role and content), agent information, model name and provider, date, task type, episode, run ID, trial name, result, verifier output, and trace source. The dataset consists of 440 training samples and is suitable for dialogue system evaluation, agent behavior analysis, and task execution result research.
提供机构:
DCAgent3


