jihuny/tldr_1.4b_10k_gopt_policy_norm_penul_last_1e-3_iter0

该数据集包含了提示文本(prompt)、索引(prompt_idx)、选择的文本(chosen)、被拒绝的文本(rejected)等字段，以及它们各自的索引和分数等信息。此外，还包括了选择的文本和被拒绝的文本的嵌入差异(embedding_difference和rm_embedding_difference)。数据集分为训练集，共有10000个样本。提供了数据集的下载大小和实际存储大小。

The dataset includes fields such as prompt, prompt_idx, chosen, rejected, and their respective indices and scores, as well as the embedding differences between the chosen and rejected texts. The dataset is split into a training set with a total of 10,000 samples. The download size and actual storage size of the dataset are provided.