divish/scratch

该数据集是通过distilabel工具创建的，包含一个`pipeline.yaml`文件，可以用于重现生成该数据集的流程。数据集的示例结构包括多个字段，如`generations`、`responses`、`instruction`、`order`、`src`、`tgt`、`domain`、`ratings`、`rationales`等。数据集的大小为n<1K，包含90个训练样本。

This dataset was created using the distilabel tool and includes a `pipeline.yaml` file that can be used to reproduce the pipeline that generated it. The datasets example structure includes multiple fields such as `generations`, `responses`, `instruction`, `order`, `src`, `tgt`, `domain`, `ratings`, `rationales`, etc. The dataset size is n<1K, containing 90 training examples.