TeichAI/MiniMax-M2.1-8800x
收藏Hugging Face2026-01-08 更新2026-02-07 收录
下载链接：
https://hf-mirror.com/datasets/TeichAI/MiniMax-M2.1-8800x
下载链接
链接失效反馈官方服务：
资源简介：
MiniMax M2.1是一个推理数据集，旨在通过微调现有的开源大型语言模型（LLMs）来创建MiniMax M2.1的蒸馏版本。数据集覆盖了多个领域，包括科学（从高中到研究生水平）、编程（通用）、网页设计、多语言创意写作、市场营销、法律、健康、金融以及一些较少见的编程语言（如Rust、Go、Ruby、Groovy等）。此外，还包含了原始的1k组通用提示（如网页开发、嵌入式系统、哲学等）。数据集的创建使用了MiniMax M2.1模型，并设置了高推理努力。数据集的统计信息显示，创建成本为42.94美元，总令牌数（输入+输出）为39.2 M。
MiniMax M2.1 is a reasoning dataset created using the MiniMax M2.1 model with high reasoning effort. It is designed for creating distilled versions of MiniMax M2.1 by fine-tuning existing open-source LLMs. The dataset covers various domains including Science (from high-school to graduate level), Programming (general), Web-Design, Multi-lingual creative writing, Marketing, Law, Health, Finance, and less common programming languages (e.g., Rust, Go, Ruby, Groovy). It also includes an original 1k set of general prompts (e.g., web-dev, embedded systems, philosophy). The dataset statistics show a cost of $42.94 (USD) and a total of 39.2 M tokens (input + output).
提供机构：
TeichAI
