snap-stanford/stark_prime_pipeline-preference_scorer_character_limit10k

该数据集是一个对话评分数据集，包含对话上下文、多个回应选项以及评分相关信息。具体包括选择的回应、被拒绝的回应、各自得分、附加信息以及相关变量。数据集分为三个子集，分别用于关系评分、文本评分和最终评分。

This dataset is a dialogue scoring dataset containing conversation contexts, multiple response options, and related scoring information. It includes chosen responses, rejected responses, their respective scores, additional information, and related variables. The dataset is divided into three subsets for relationship scoring, text scoring, and final scoring.