DCAgent2/eval-DCAgent_a1-agenttuning_mind2web_DCAgent2_swebench-verified-random-100-folders
收藏资源简介:
该数据集是一个多任务对话数据集,包含AI代理或模型之间的交互记录。数据特征包括对话内容(conversations,含内容和角色字段)、代理(agent)、模型(model)、模型提供方(model_provider)、日期(date)、任务(task)、事件(episode)、运行ID(run_id)、试验名称(trial_name)、结果(result)、验证器输出(verifier_output)和追踪来源(trace_source)。数据集仅提供训练划分(train),共6546个示例,总大小约为868.9 MB,适用于自然语言处理任务,如对话分析、模型评估或多代理系统研究。
This dataset is a multi-task dialogue dataset containing interaction records between AI agents or models. The features include conversations (with content and role fields), agent, model, model provider, date, task, episode, run ID, trial name, result, verifier output, and trace source. The dataset provides only a training split (train) with 6,546 examples and a total size of approximately 868.9 MB, suitable for natural language processing tasks such as dialogue analysis, model evaluation, or multi-agent system research.




