Bmingg/DPO_6b_llama_3600_best_constrained_bleurt_vs_ancestral_sampling_bleurt_strict_bleurt

该数据集包含三个字段：提示(prompt)、选中(chosen)和拒绝(rejected)，均为字符串类型。数据集仅包含训练集部分，共有3600个示例，数据大小为1782361字节。数据集适用于可能需要根据提示选择或拒绝某些内容的场景。The dataset includes three fields: prompt, chosen, and rejected, all of which are string types. The dataset contains only the training set with a total of 3600 examples, with a size of 1782361 bytes. The dataset is suitable for scenarios where content may need to be selected or rejected based on a prompt.