classified-embeddings
收藏arXiv2025-09-30 收录
下载链接：
https://api.semanticscholar.org/api-docs/datasets
下载链接
链接失效反馈官方服务：
资源简介：
该数据集将文档嵌入向量与来自分类论文数据集的元数据和分类信息相关联。文档嵌入向量是通过使用在SciBERT上微调的SPECTER2模型生成的。该数据集包含了134,705,300个文档嵌入向量，其任务是对各领域使用文档嵌入分析语义变化。
This dataset associates document embeddings with metadata and classification information sourced from a categorized academic paper dataset. The document embeddings are generated using the SPECTER2 model fine-tuned on SciBERT. This dataset contains 134,705,300 document embeddings, and its core task is to analyze semantic changes across various domains using these embeddings.
提供机构：
Semantic Scholar
