distilabel-internal-testing/embeddings-dataset-semantically-similar

该数据集包含一个`pipeline.yaml`文件，可用于在distilabel中复现生成数据集的流程。数据集的结构包括一个名为default的配置示例，展示了数据的JSON格式和加载方法。数据示例包含anchor、distilabel_metadata、model_name、negative和positive字段，分别表示锚点文本、元数据、模型名称、负面文本和正面文本。

原始信息汇总

数据集概述

数据集基本信息

• 名称: embeddings-dataset-paraphrase
• 创建工具: distilabel
• 大小分类: 小于1000条记录
• 标签:
  • synthetic
  • distilabel
  • rlaif

数据集结构

• 配置: 默认
• 数据示例结构: json { "anchor": "Astrology: I am a Capricorn Sun Cap moon and cap rising...what does that say about me?", "distilabel_metadata": { "raw_output_paraphrase": "## Positive

With a triple Capricorn influence, youre likely a driven and ambitious individual with a strong sense of discipline and responsibility.

Negative

The cap on my pen is always getting lost, and its really frustrating when I need to sign important documents." }, "model_name": "meta-llama/Meta-Llama-3-70B-Instruct", "negative": "The cap on my pen is always getting lost, and its really frustrating when I need to sign important documents.", "positive": "With a triple Capricorn influence, youre likely a driven and ambitious individual with a strong sense of discipline and responsibility." }

数据集加载

• 加载方式: python from datasets import load_dataset

  ds = load_dataset("distilabel-internal-testing/embeddings-dataset-paraphrase", "default")

  或简化为: python from datasets import load_dataset

  ds = load_dataset("distilabel-internal-testing/embeddings-dataset-paraphrase")