Simonkami/MctsAI_selftrain_fireball

该数据集是一个本地混合的监督微调（SFT）数据集，由两部分组成：原始数据集`Simonkami/MctsAi_Zoning`的`train`分割和自训练文件`selftrain_good_samples_round1.jsonl`。混合设置中，原始数据集使用了2000条样本（占总原始数据的20%），自训练数据集使用了4192条样本，最终总样本数为6192条。数据集的列包括`instruction`、`input`和`output`。`train_with_meta.jsonl`文件包含额外的`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/MctsAi_Zoning` split `train` and the self-training file `selftrain_good_samples_round1.jsonl`. The mixing settings include using 2000 samples from the original dataset (20% of the total original data) and 4192 samples from the self-training dataset, resulting in a final total of 6192 samples. The dataset columns include `instruction`, `input`, and `output`. The file `train_with_meta.jsonl` contains extra `source` and `score` fields, but `train.jsonl` is the recommended file for SFT training.