saurabh5/too-big-stuff

这是一个用于对话或角色扮演任务的数据集，包含消息内容、角色信息、真实标签、唯一标识符、数据源标识、自定义ID、原始数据集标识等信息。数据集还提供了总回合数、正确回合数和通过率等评估指标。

This is a dataset for dialogue or role-playing tasks, containing message content, role information, ground truth labels, unique identifiers, dataset source identifiers, custom IDs, original dataset identifiers, etc. The dataset also provides evaluation metrics such as total rollouts, correct rollouts, and passrate.