AlexHung29629/stack-exchange-paired-128K

该数据集名为stack-exchange-paired-128K，包含128,000个训练示例，每个示例由prompt、chosen和rejected三个字符串特征组成。数据集总大小为243,412,260字节，下载大小为82,603,750字节。数据集标签为dpo，与llama2模型的token数有关，具体为97,868,021个token。

The dataset named stack-exchange-paired-128K contains 128,000 training examples, each consisting of three string features: prompt, chosen, and rejected. The total size of the dataset is 243,412,260 bytes, with a download size of 82,603,750 bytes. The dataset is tagged with dpo and is related to the llama2 models token count, specifically 97,868,021 tokens.

原始信息汇总

数据集卡片 "stack-exchange-paired-128K"

数据集信息

特征

• prompt: 数据类型为字符串
• chosen: 数据类型为字符串
• rejected: 数据类型为字符串

分割

• train:
  • 字节数: 243412260
  • 样本数: 128000

下载和数据集大小

• 下载大小: 82603750 字节
• 数据集大小: 243412260 字节