plaguss/test_personahub_fineweb_edu

该数据集是通过distilabel工具创建的，包含一个pipeline.yaml文件，用于复现生成数据集的流程。数据集的每个样本包含text、persona、distilabel_metadata和model_name四个字段。其中，text字段包含一段文本，persona字段描述了文本的生成者角色，distilabel_metadata字段包含生成文本的原始输出，model_name字段指定了生成文本所使用的模型。数据集只有一个配置（default），并且可以通过Hugging Face的datasets库加载。

This dataset was created using the distilabel tool and includes a pipeline.yaml file to reproduce the pipeline that generated it. Each example in the dataset contains four fields: text, persona, distilabel_metadata, and model_name. The text field contains a piece of text, the persona field describes the role of the text generator, the distilabel_metadata field contains the raw output of the generated text, and the model_name field specifies the model used to generate the text. The dataset has only one configuration (default) and can be loaded using the Hugging Face datasets library.

原始信息汇总

数据集概述

数据集结构

特征

• text: 类型为字符串。
• persona: 类型为字符串。
• distilabel_metadata: 结构化数据，包含以下字段：
  • raw_output_text_to_persona_0: 类型为字符串。
• model_name: 类型为字符串。

数据分割

• train: 包含3个样本，占用14689字节。

配置

• default: 包含一个数据文件，路径为data/train-*。

标签

• synthetic
• distilabel
• rlaif

数据集大小

• 下载大小: 13811字节
• 数据集大小: 14689字节

示例

json { "distilabel_metadata": { "raw_output_text_to_persona_0": "A health-conscious food blogger or a specialty tea shop owner knowledgeable about tea varieties and their benefits." }, "model_name": "meta-llama/Meta-Llama-3-70B-Instruct", "persona": "A health-conscious food blogger or a specialty tea shop owner knowledgeable about tea varieties and their benefits.", "text": "Ever wondered what white tea is? Why we call it white anyway? Here are a few interesting facts about white tea, made from the plant Camellia sinensis: - White tea is similar to green tea in that it has undergone little processing and no fermentation whatsoever. - The leaves are harvested before they open fully, when the unopened buds are still covered by thin silvery white hair. Hence the white tea name. - White tea has a subtle and slightly sweet flavor that sharpens the senses without overwhelming the tongue, whereas green tea has a u201cgrassyu201d taste. - There is 25% to 30% less caffeine in white tea than in green tea, though it depends on the variety. - White tea health benefits are the highest among the tea familyu00b9. - It is more expensive than green or black tea because white tea is scarcer. Yet, at Savorique, white tea is as cheap as black or green tea. Note 1: Science Daily u201cu201dWhite Tea Could Keep You Healthy and Looking Youngu201d", (August 14, 2009) and Science Daily u201cu201dWhite Tea Beats Green Tea In Fighting Germsu201d", (May 28, 2004)." }

加载数据集

python from datasets import load_dataset

ds = load_dataset("plaguss/test_personahub_fineweb_edu")