DCAgent/g1_min_episodes_e1_weighted_issue_20k_glm47_traces
收藏资源简介:
该数据集包含多轮对话数据,用于AI代理或任务执行场景。每个数据样本包括对话内容(conversations,其中包含角色和文本内容)、代理类型(agent)、模型信息(model和model_provider)、日期(date)、任务类型(task)、事件集(episode)、运行标识(run_id)、试验名称(trial_name)、结果(result)、验证器输出(verifier_output,当前为空)和来源追踪(trace_source)。数据集适用于自然语言处理、对话系统训练或AI行为分析,总大小为约468MB,包含13783个训练样本。
This dataset contains multi-turn conversation data for AI agent or task execution scenarios. Each sample includes conversation content (with roles and text), agent type, model information (model and model provider), date, task type, episode, run ID, trial name, result, verifier output (currently null), and trace source. It is suitable for natural language processing, dialogue system training, or AI behavior analysis, with a total size of approximately 468MB and 13,783 training examples.




