allenai/Dolci-Think-RL-7B-Completions-SFT
收藏Hugging Face2025-12-12 更新2025-12-20 收录
下载链接：
https://hf-mirror.com/datasets/allenai/Dolci-Think-RL-7B-Completions-SFT
下载链接
链接失效反馈官方服务：
资源简介：
Dolci-Think-Completions-SFT是一个包含5,031,398个来自Olmo-3-7B-Think-SFT模型补全的数据集，主要用于过滤简单数据。数据集包含636,095个高质量提示，覆盖数学、代码、精确指令遵循、通用聊天和谜题五个领域。数据来源多样，包括IF Multi-Constraint、OMEGA Math、AceCoder等多个数据集，并经过关键词和主题过滤、基于执行的测试案例验证、F1分数过滤等多种处理。数据集分为coding、general、if、math和puzzle五个分割，适用于研究和教育用途。
Dolci-Think-Completions-SFT is a set of 5,031,398 completions from the Olmo-3-7B-Think-SFT model over the prompts considered when making Dolci-Think-RL. These completions were mainly used to filter easy data. It contains 636,095 high-quality prompts covering Math, Code, Precise Instruction Following, General Chat, and Puzzles. The dataset is composed of multiple sources including IF Multi-Constraint, OMEGA Math, AceCoder, etc., and has undergone various processing such as keyword & topic filtering, execution-based test-case validation, and F1-score filtering. The dataset is split into coding, general, if, math, and puzzle, and is licensed under ODC-BY for research and educational use.
提供机构：
allenai
