BabiStories
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/facebookresearch/MemoryMosaics
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一组为语言建模实验而生成的小故事集合，其内容设计得足够简单，以至于一个三岁的孩子都能理解。该数据集名为BabiStories，是使用Mixtral-8x7B开放语言模型生成的，旨在提高生成文本中的叙事连贯性和语言质量。其任务目标是语言建模。
This dataset is a collection of short stories generated for language modeling experiments, with content designed to be simple enough for even a three-year-old child to understand. Named BabiStories, it was generated using the Mixtral-8x7B open language model, aiming to improve the narrative coherence and linguistic quality of the generated text. Its task objective is language modeling.
提供机构：
Facebook Research
