llamastack/ruler

这个数据集包含了对话信息以及预期的回答。每个样本由一个唯一的标识符、对话内容（包括发言内容和角色）以及预期的答案组成。数据集分为多个配置，包括cwe、fwe、qa_1、qa_2和vt，每个配置都有对应的测试集和验证集，用于不同的训练和验证目的。

This dataset consists of dialogues and expected answers. Each sample includes a unique identifier, dialogue content (including speech content and role), and the expected answer. The dataset is divided into multiple configurations, including cwe, fwe, qa_1, qa_2, and vt, each with corresponding test and validation sets for different training and validation purposes.