ljvmiranda921/gsd-smith-Japanese
收藏资源简介:
该数据集包含3058个训练示例,主要用于对话或代理交互任务,具有结构化特征,如id(唯一标识符)、seed_prompt(种子提示)、language(语言类型)、model(模型名称)、messages(包含角色和内容的对话消息列表)、agent_trace(代理交互轨迹的JSON列表)、research_early_stopping(研究早期停止标志)和source_id(来源标识)。数据集仅提供训练划分,下载大小约为80.05 MB,总数据集大小约为81.27 MB。
This dataset contains 3058 training examples, primarily designed for dialogue or agent interaction tasks, with structured features including id (unique identifier), seed_prompt (seed prompt), language (language type), model (model name), messages (a list of dialogue messages with role and content), agent_trace (a JSON list of agent interaction traces), research_early_stopping (research early stopping flag), and source_id (source identifier). The dataset only provides a train split, with a download size of approximately 80.05 MB and a total dataset size of approximately 81.27 MB.




