WPRM/preference_data_version_1

该数据集包含了一系列任务交互的相关信息。每个交互包括选中的(chosen)和拒绝的(rejected)选项，每个选项有内容(content)和角色(role)信息。此外，还包括任务索引(task_idx)、动作索引(action_idx)、意图(intent)、轨迹(trajectory)、当前URL(current_url)、文本观察(text_observation)和清单(checklist)。数据集仅包含训练集划分，共有1810个示例，总文件大小为62862221字节。

The dataset consists of a series of task interaction information. Each interaction includes chosen and rejected options, with each option having content and role information. Additionally, it includes task index (task_idx), action index (action_idx), intent (intent), trajectory (trajectory), current URL (current_url), text observation (text_observation), and checklist (checklist). The dataset is split only into a training set with a total of 1810 examples and a total file size of 62862221 bytes.