匠数科技大模型sft数据集

<p>匠数大模型SFT数据集由匠数科技整理，包含1138万条中文和276万条英文高质量SFT数据，涵盖50类任务。数据经统一格式化、清洗及严格内容审核，确保安全可靠。提供中英文类别关键词，适用于大模型指令微调、多任务学习与内容安全研究，支持开放学术与产业应用。</p>

The Jiangshu Large Model SFT Dataset, curated by Jiangshu Technology, consists of 11.38 million high-quality Chinese SFT instances and 2.76 million high-quality English SFT instances, covering 50 task categories. All data has undergone unified formatting, cleaning and strict content auditing to ensure security and reliability. It provides category keywords in both Chinese and English, and is applicable to instruction fine-tuning of large language models (LLMs), multi-task learning and content security research, supporting open academic and industrial applications.