bethgelab/sober_reasoning
收藏Hugging Face2025-04-11 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/bethgelab/sober_reasoning
下载链接
链接失效反馈官方服务:
资源简介:
Sober Reasoning数据集是一个包含了不同硬件集群上模型评估日志的数据集,用于分析和评估语言模型在推理任务中的表现。数据集涵盖了多个基准测试和多种类型的模型,包括基于RL的模型、基于SFT的模型和基线模型。
The Sober Reasoning dataset consists of evaluation logs and outputs from different hardware clusters, used for analyzing and assessing the performance of language models in reasoning tasks. The dataset covers multiple benchmark tests and various types of models, including RL-based models, SFT-based models, and baseline models.
提供机构:
bethgelab



