Simonkami/bvb_selftrain2

该数据集是一个本地混合的SFT数据集，由原始数据集`Simonkami/bvb_8000`的`train`分割和自训练文件`selftrain_good_samples_round1.jsonl`构建而成。混合设置包括原始数据集和自训练数据集的使用数量、比例等。数据集包含`instruction`、`input`和`output`三列，`train_with_meta.jsonl`还包含额外的`source`和`score`字段，但推荐使用`train.jsonl`进行SFT训练。

This dataset is a local mixed SFT dataset built from the original dataset `Simonkami/bvb_8000` split `train` and the self-training file `selftrain_good_samples_round1.jsonl`. The mixing settings include the number of samples used from the original and self-training datasets, their ratios, etc. The dataset contains columns `instruction`, `input`, and `output`. `train_with_meta.jsonl` includes additional `source` and `score` fields, but `train.jsonl` is recommended for SFT training.