flowaicom/Feedback-Bench

Feedback-Bench数据集是一个专门用于评估不同标准和量表下的语言模型评分器的克隆数据集。它包含评分描述、原始指令、输出、响应、参考答案、反馈和评分标准等字段，并且已经添加了唯一的标识符列。Feedback-Bench dataset is a clone dataset specialized in benchmarking LM judges across different criteria and rubrics. It includes fields such as scoring descriptions, original instructions, outputs, responses, reference answers, feedback, and scoring criteria, and a unique identifier column has been added.