v1ctor10/BERT_SBERT_embeddings_SAE

该数据集包含多个字段，涉及公司标识符、年份、公司名称、行业分类代码、输入标识符、股票代码、回报率、月度回报率矩阵、索引级别、输入标识符长度、行业分类以及BERT和SBERT的嵌入向量。数据集被分割为训练集，包含26,769个样本，总大小为2,519,325,898字节，下载大小为947,775,283字节。

This dataset includes multiple fields such as company identifier (cik), year, company name, industry classification code (sic_code), input identifiers (input_ids), ticker symbols, returns, logged monthly returns matrix, index level (__index_level_0__), input identifiers length, industry classification, and BERT and SBERT embeddings. The dataset is split into a training set containing 26,769 samples, with a total size of 2,519,325,898 bytes and a download size of 947,775,283 bytes.