jihuny/tldr_1.4b_10k_gopt_policy_unnorm_penul_last_1e-6_iter0

该数据集包含了文本对的比较信息，每条记录包括一个提示文本（prompt）和两个候选文本（chosen和rejected），以及它们的相关索引、分数、对数概率和完成原因等。数据集分为训练集，共有10000个样本。

The dataset consists of comparison information for text pairs, each record includes a prompt text and two candidate texts (chosen and rejected), along with their respective indices, scores, log probabilities, and completion reasons, etc. The dataset is split into a training set with a total of 10,000 samples.