sentaibouzha/lima-qwen25-pairrm-preference-dataset

该数据集包含50个训练样本，每个样本由提示文本、采纳的回复文本和拒绝的回复文本组成，用于支持基于对比学习的对话模型训练或评估。数据集大小为72,122字节，下载大小为55,359字节，特征包括prompt（字符串类型）、chosen（字符串类型）和rejected（字符串类型）。

This dataset contains 50 training samples, each consisting of a prompt text, a chosen response text, and a rejected response text, designed to support dialogue model training or evaluation based on contrastive learning. The dataset size is 72,122 bytes with a download size of 55,359 bytes, and features include prompt (string type), chosen (string type), and rejected (string type).