selfcorrexp2/llama3_sft_balanced_rr60k_train_on_corr_ep3tmp10_vllmexp
收藏Hugging Face2025-01-07 更新2025-04-26 收录
下载链接：
https://hf-mirror.com/datasets/selfcorrexp2/llama3_sft_balanced_rr60k_train_on_corr_ep3tmp10_vllmexp
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含多个字段，如索引(idx)，提示(prompt)，答案(answers)，真实标签(gt)，代理标签(proxy_label)，以及第二轮奖励(second_round_rewards)。数据集分为训练集(train)，共有1000个样本。然而，README文件中并未提供数据集的具体中文描述。
The dataset includes fields such as index (idx), prompt, answers, ground truth (gt), proxy label (proxy_label), and second round rewards (second_round_rewards). The dataset is split into a training set (train) with a total of 1000 samples. However, the README file does not provide a specific English description of the dataset.
提供机构：
selfcorrexp2
