jihuny/tldr_1.4b_10k_random_global_unnorm_penul_mean_iter0

这是一个包含文本及其相关信息的的数据集，具体字段包括提示文本(prompt)、提示索引(prompt_idx)、选中的文本(chosen)、被拒绝的文本(rejected)、选中/被拒绝文本的索引、分数、对数概率、完成原因以及嵌入向量差异等。数据集分为训练集(train)，共有10000个示例，总大小约为171MB。提供了默认配置以指定训练数据文件的路径。

This dataset contains text and its related information, including fields such as prompt text, prompt index, chosen text, rejected text, indices of chosen/rejected texts, scores, log probabilities, finish reasons, and embedding differences. The dataset is split into a training set (train) with a total of 10,000 examples and a total size of approximately 171MB. A default configuration is provided to specify the path to the training data files.