DataPilot/Zero_SFT_Ja_v3_Reasoning
收藏Hugging Face2025-05-11 更新2025-07-05 收录
下载链接:
https://hf-mirror.com/datasets/DataPilot/Zero_SFT_Ja_v3_Reasoning
下载链接
链接失效反馈官方服务:
资源简介:
这是一个包含高质量合成提示及其AI输出的日本语数据集,所有数据都是使用Qwen3-235B-A22B模型从零开始合成的。数据集采用JSONL格式,适用于日语对话型AI长考模型的微调、命令遵守型模型的基准测试以及提示工程研究和评估。
This dataset contains high-quality synthetic prompts and their AI outputs in Japanese, with all data synthesized from scratch using the Qwen3-235B-A22B model. The dataset is in JSONL format and is suitable for fine-tuning of Japanese conversational AI long-thinking models, creating benchmarks for instruction-following models, and prompt engineering research and evaluation.
提供机构:
DataPilot



