LLLMs Survey Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/a-kostikova/LLLMs-Survey
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一组从ACL和arXiv论文语料库中提取的、专注于大型语言模型局限性的标注摘要。它包含了针对以大型语言模型为重点的论文的评分，这些评分从0到5不等，用以表示对这些局限性的讨论深度。在规模上，该数据集爬取了25万篇论文，并从中提取了14,648篇关于局限性的论文。该数据集的任务是对大型语言模型研究论文中的局限性进行分类和证据提取。
This dataset is a collection of annotated abstracts focusing on the limitations of large language models (LLMs), extracted from the ACL and arXiv paper corpora. It includes scores ranging from 0 to 5 assigned to papers centered on LLMs, which reflect the depth of their discussion on these limitations. In terms of scale, the dataset crawled 250,000 papers and extracted 14,648 papers that discuss the limitations of LLMs. The task of this dataset is to classify the limitations in LLM-focused research papers and extract supporting evidence for such limitations.
提供机构：
A. Kostikova et al.
