DCAgent3/gaia_127_rl__24GPU_shaped__exp_rpt_pymethods2test_large__GLM_4_7_swesmith_san_3ab1d39d3
收藏资源简介:
该数据集是一个包含多轮对话和任务执行记录的数据集,主要用于分析和评估代理或模型在交互任务中的表现。数据集包含401个训练示例,每个示例包括对话内容(conversations,其中role表示角色如用户或助手,content表示对话文本)、代理标识(agent)、模型信息(model和model_provider)、日期(date)、任务类型(task)、情节(episode)、运行ID(run_id)、试验名称(trial_name)、结果(result)、验证器输出(verifier_output)和跟踪来源(trace_source)等字段。这些字段均为字符串类型,数据集大小为约10.37 MB,适用于自然语言处理任务,如对话系统分析、模型性能评估或任务自动化研究。
This dataset contains multi-turn conversations and task execution records, primarily designed for analyzing and evaluating the performance of agents or models in interactive tasks. It includes 401 training examples, each featuring fields such as conversations (with role indicating the participant like user or assistant, and content as the dialogue text), agent identifier, model information (model and model_provider), date, task type, episode, run ID, trial name, result, verifier output, and trace source, all stored as string types. The dataset size is approximately 10.37 MB, suitable for natural language processing applications, including dialogue system analysis, model performance evaluation, or task automation research.




