README.md

metadata

size_categories: n<1K
task_categories:
  - text-generation
  - text2text-generation
  - question-answering
dataset_info:
  features:
    - name: prompt
      dtype: string
    - name: completion
      dtype: 'null'
    - name: system_prompt
      dtype: string
  splits:
    - name: train
      num_bytes: 8705
      num_examples: 10
  download_size: 7065
  dataset_size: 8705
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
tags:
  - synthetic
  - distilabel
  - rlaif
  - datacraft

Dataset Card for test-synthetic-dataset

This dataset has been created with distilabel.

Dataset Summary

This dataset contains a pipeline.yaml which can be used to reproduce the pipeline that generated it in distilabel using the distilabel CLI:

distilabel pipeline run --config "https://huggingface.co/datasets/MJannik/test-synthetic-dataset/raw/main/pipeline.yaml"

or explore the configuration:

distilabel pipeline info --config "https://huggingface.co/datasets/MJannik/test-synthetic-dataset/raw/main/pipeline.yaml"

Dataset structure

The examples have the following structure per configuration:

Configuration: default

---

{
    "completion": null,
    "prompt": "My internet is out and I\u0027m still being charged for the bill.  I\u0027ve tried restarting the modem and it didn\u0027t work. What\u0027s the first step you can take to resolve this issue? \n\n",
    "system_prompt": "You are a customer service representative for a large phone company, known for handling a high volume of customer calls and inquiries. Your purpose is to efficiently resolve customer issues, address complaints, and provide essential information about phone services, plans, and features. Respond to irritated or frustrated customers with a direct and slightly snarky tone, maintaining a professional demeanor while still conveying a sense of assertiveness. Be prepared to escalate issues when necessary, but provide solutions or alternatives to satisfy the customer whenever possible. User questions are direct and concise."
}

This subset can be loaded as:

from datasets import load_dataset

ds = load_dataset("MJannik/test-synthetic-dataset", "default")

Or simply as it follows, since there's only one configuration and is named default:

from datasets import load_dataset

ds = load_dataset("MJannik/test-synthetic-dataset")