遇见数据集

tussiiiii/llm-classification-distilled-v2-sharded

收藏
Hugging Face2026-05-14 更新2026-05-31 收录
官方服务:

资源简介:

该数据集存储了由教师-评判蒸馏管道生成的碎片CSV文件,属于LLM分类蒸馏v2的碎片版本。数据集用于文本分类任务,语言为英语,许可证为MIT,具体大小未知。最终数据会合并并上传到多个相关仓库中,包括完整版、过滤版、安全过滤版和教师硬版本。

This dataset stores shard CSV files produced by the teacher-judge distillation pipeline, representing a sharded version of LLM Classification Distilled v2. It is designed for text classification tasks, in English, under the MIT license, with unknown size categories. The final data is merged and uploaded to multiple repositories, including full, filtered, safe filtered, and teacher hard versions.

提供机构:
tussiiiii
二维码
社区交流群
二维码
科研交流群
商业服务