Dataset for "A 5/3 spectral law in contextual language representations"

This repository contains the dataset and analysis code for our research on spectral properties of word embeddings from human-written and AI-generated texts across multiple languages.本代码仓库收录了我们针对多语种人工撰写文本与AI生成文本的词嵌入（word embeddings）频谱特性展开研究的相关数据集与分析代码。