daikin-industries-ltd/ja-fineweb-2-hvac-fastText-scored-v6
收藏数据链接：
官方服务：
资源简介：
本数据集是基于日语空調（HVAC）相关文本数据，通过FastText分类模型和LLM进行质量评分后的结果。数据集包含了文本内容、来源URL、FastText分类分数（0.0-1.0）、LLM质量分数（1-5）及其评分依据。数据集筛选了FastText分数较高的文档，并进行了LLM的详细质量评估。数据规模为200,000条记录，主要用于空調技术人员的教育用途。
This dataset is based on Japanese HVAC-related text data, with quality scores added using FastText classification model and LLM. It includes text content, source URL, FastText classification score (0.0-1.0), LLM quality score (1-5), and the rationale for the score. The dataset extracts only documents with high FastText scores and conducts detailed quality evaluation by LLM. The data scale is 200,000 records, mainly for educational purposes of HVAC technicians.
