EleutherAI/reasoning-mix
收藏Hugging Face2025-01-24 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/EleutherAI/reasoning-mix
下载链接
链接失效反馈官方服务：
资源简介：
这是一个混合型数据集，包含高质量的网页文本、数学推理以及数学、编程、科学和谜题数据，用于高级推理模型（如解耦自编码器、转码器）的训练。数据集由训练集组成，共有11693047个样本，总大小约为55.7GB。
This is a mixed dataset containing high-quality web text, math reasoning, as well as math, coding, science, and puzzle data, intended for training advanced reasoning models such as disentangled autoencoders and transcoders. The dataset consists of a training set with a total of 11,693,047 samples, totaling approximately 55.7GB in size.
提供机构：
EleutherAI
