JinHyeong777/Llama-3.2-1B-Instruct-best_of_n-prm-completions
收藏Hugging Face2025-02-11 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/JinHyeong777/Llama-3.2-1B-Instruct-best_of_n-prm-completions
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含数学问题的数据集，每个问题都有相应的解决方案、答案和预测结果。数据集按不同的种子、聚集策略和解码策略分为多个配置，每个配置下都有500个样本的训练集和1个样本的评估集。
This is a dataset containing math problems, each with corresponding solutions, answers, and prediction results. The dataset is divided into multiple configurations based on different seeds, aggregation strategies, and decoding strategies, with each configuration having a training set of 500 samples and an evaluation set of 1 sample.
提供机构：
JinHyeong777
