geniacllm/CulturaY-ja-askllm-deduped-v1
收藏Hugging Face2024-08-11 更新2024-06-12 收录
下载链接：
https://hf-mirror.com/datasets/geniacllm/CulturaY-ja-askllm-deduped-v1
下载链接
链接失效反馈官方服务：
资源简介：
CulturaY-ja-askllm-deduped-v1数据集是多语言数据集CulturaY的日语部分，通过Ask-LLM方法进行评分并去重。数据集在原有列的基础上增加了askllm_score列，用于存储Ask-LLM的评分。Ask-LLM评分使用了RakutenAI-7B-instruct模型，并提供了具体的评分提示。
This dataset is the Japanese part of the multilingual dataset ontocord/CulturaY, scored using the Ask-LLM method, with an additional column askllm_score. After scoring, duplicate removal was performed. The LLM model used is Rakuten/RakutenAI-7B-instruct, with scoring criteria including the format of the data point, knowledge content, and whether it contains harmful content.
提供机构：
geniacllm
