selfcorrexp/llama3_regular_NON_balanced_sft_4_ORM_training
收藏Hugging Face2024-12-29 更新2025-02-15 收录
下载链接：
https://hf-mirror.com/datasets/selfcorrexp/llama3_regular_NON_balanced_sft_4_ORM_training
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含对话数据的训练集，数据集的字段包括索引、真实标签、提示信息、难度等级、类型、解决方案、预测结果、对话轮数、对话内容（包括角色和内容）、是否为第一轮、奖励标识和标记位。数据集总共包含326982个训练样本，大小为625265293字节。
This is a training dataset containing conversational data, with fields including index, ground truth, prompt, difficulty level, type, solution, prediction, number of turns in conversation, conversation content (including role and content), whether it is the first round, reward indication, and flag. The dataset contains a total of 326982 training samples, with a size of 625265293 bytes.
提供机构：
selfcorrexp
