kajuma/Llama-SFT-3000

该数据集包含用于文本生成任务的日语数据，主要由两个特征组成：prompt和chosen。数据集包含3000个训练样本，总大小为8797671字节。数据集的指令由Tanuki-8B-dpo-v1.0生成，输出由Llama-3.1-SuperSwallow-70B-Instruct-v0.1生成。Llama 3.1的许可证是Llama 3.1 Community License。

This dataset includes two main features: prompt and chosen, both of which are string type. The dataset is divided into a training set, containing 3000 samples. The language of the dataset is Japanese, with a size between 1K and 10K. The generation instructions of the dataset are produced by the Tanuki-8B-dpo-v1.0 model, and the outputs are generated by the Llama-3.1-SuperSwallow-70B-Instruct-v0.1 model. The dataset is licensed under the Llama 3.1 Community License.