Nexdata/100000_Fine-Tuning_text_data_set_for_Korean_LLM_General_Domain_SFT

这是一个包含10万条韩语微调文本数据集，用于LLM通用领域的训练。该数据集包含12种不同类型的SFT QA，准确率不低于95%。所有提示均为手动编写，以确保覆盖多样性。数据集的类别包括头脑风暴、聊天、分类、封闭式问答、代码、提取、生成、开放式问答、推理、改写、总结等。数据格式为xlsx。

This is a fine-tuned text dataset with 100,000 Korean entries for LLM general domain training. It contains 12 types of SFT QA with an accuracy of not less than 95%. All prompts are manually written to ensure diversity coverage. The dataset categories include brainstorming, chat, classification, closed QA, code, extract, generation, open QA, reason, rewrite, summarization, etc. The data format is xlsx.