gogo8232/experiment_first_half
收藏Hugging Face2024-07-11 更新2024-07-22 收录
下载链接:
https://hf-mirror.com/datasets/gogo8232/experiment_first_half
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,如指令(instruction)、输入(input)、输出(output)等,主要用于处理与指令相关的数据。数据集还包括指令困惑度(instruction_perplexity)、困惑度(perplexity)、合成响应(synthetic_response)、偏好(preference)、原因(reason)和精选指令(curated_instruction)等字段,这些字段可能用于评估或生成指令的响应。数据集包含一个训练分割(train),共有17431个例子,总大小为39410131字节。
This dataset includes multiple fields such as instruction, input, output, etc., primarily used for handling instruction-related data. The dataset also includes fields like instruction perplexity, perplexity, synthetic response, preference, reason, and curated instruction, which may be used for evaluating or generating responses to instructions. The dataset contains a training split (train) with 17431 examples, totaling 39410131 bytes.
提供机构:
gogo8232
原始信息汇总
数据集概述
数据集信息
特征
- instruction: 字符串类型
- input: 字符串类型
- output: 字符串类型
- instruction_perplexity: 浮点数类型
- perplexity: 浮点数类型
- synthetic_response: 字符串类型
- preference: 整数类型
- reason: 字符串类型
- curated_instruction: 字符串类型
数据分割
- train:
- 字节数: 39410131
- 样本数: 17431
数据集大小
- 下载大小: 21306641 字节
- 数据集大小: 39410131 字节
配置
- config_name: default
- data_files:
- split: train
- path: data/train-*
- data_files:



