JakeOh/iself-sft_gt-gsm8k-llama1b

该数据集包含多个特征，包括文档哈希值、问题、真实答案、模型生成的答案、对话消息列表以及一个布尔值表示答案是否完全匹配。数据集分为训练集和测试集，分别包含50898和8653个样本。数据集的总下载大小为69027866字节，总大小为143831425字节。

The dataset contains multiple features, including document hash (doc_hash), question, ground truth answer (gt_answer), model-generated answer (answer), a list of conversation messages (messages), and a boolean value (exact_match) indicating whether the answer is an exact match. The dataset is divided into training and test sets, containing 50898 and 8653 samples respectively. The total download size of the dataset is 69027866 bytes, and the total size is 143831425 bytes.