UCLAML/synthetic_data_mistral-7b-instruct-sppo-iter1_score

该数据集名为synthetic_data_mistral-7b-instruct-sppo-iter1_score，主要用于训练和测试模型。数据集包含chosen和rejected两部分，每部分都有content和role两个字段，数据类型为字符串。此外，数据集还包含三个浮点型字段：chosen_probs、chosen_probs_win和chosen_probs_lose。数据集分为训练集和测试集，训练集有19766个样本，测试集有500个样本。数据集的下载大小为50796684字节，总大小为98584369字节。

The dataset named synthetic_data_mistral-7b-instruct-sppo-iter1_score is primarily used for training and testing models. It includes chosen and rejected sections, each with content and role fields, both of string type. Additionally, the dataset contains three float64 fields: chosen_probs, chosen_probs_win, and chosen_probs_lose. The dataset is divided into a training set with 19,766 samples and a test set with 500 samples. The download size of the dataset is 50,796,684 bytes, and the total size is 98,584,369 bytes.