metadata

license: cc-by-sa-3.0
language:
  - en
task_categories:
  - text-generation
  - fill-mask
tags:
  - language-modeling
  - masked-language-modeling
pretty_name: SuperWIKI Cleaned

Dataset Card for SuperWIKI Cleaned

Dataset Description

Homepage: (TODO)
Repository: N/A
Paper: N/A
Leaderboard: N/A
Point of Contact: KaraKaraWitch

Dataset Summary

If you show most of those to people and ask them to form an opinion,
the answer isn't just going to be "I don't know": it'll be "I don't care."

Tom Scott

SuperWIKI Cleaned is a focused dataset on wikipedia articles. This dataset is derived from raw files provided in SuperWIKI.

Supported Tasks and Leaderboards

The dataset is generally used for Language Modeling.

Languages

English

Dataset Structure

All the files are located in gzip'd jsonl files.

Data Instances

Refer to this sample to see all the fields:

{
    "text": "|Abbreviation||WAY|\n|Formation||1966|\n|Headquarters||Wat Ananda Metyarama Thai Buddhist Temple|\nPresident\n|Phanom Sinth Suwanarat|\n|Affiliations||Wat Ananda Metyarama Thai Buddhist Temple|\n|Website||http://way.org.sg|\nThe\n**Wat Ananda Youth** is a Buddhist youth organisation... <TRUNCATED>",
    "title": "Wat Ananda Youth",
    "url": "https://en.wikipedia.org/wiki/Wat_Ananda_Youth",
    "issues": [
        "template:original research"
    ],
    "selectors": [
        "table.box-Original_research"
    ],
    "templates": [
        "template:reflist",
        "template:infobox organization",
        "template:cite web",
        "template:original research",
        "template:main"
    ]
}

Data Fields

text: The HTML Text (After pre-processing) from SuperWIKI converted to markdown with links removed and formatting kept.
title: The title of the wikipedia article.
issues: A list of custom list of templates that has been removed from the html (ala, pre-processing) for the article.
selectors: issues are based on templates, which may have multiple templates but mean the same thing.
In that case, the selectors provide a unduplicated css class selectors that were used for the article. (Template:Few sources is the same as Template:More citations needed for example.)
templates: Used for debugging but are all the templates found in the article.

Q-Score Distribution

Not Applicable

Data Splits

No data splits were done.

Dataset Creation

Curation Rationale

"Wikipedia is a wonderful resources however it could be considered too sparse as there are many articles that are not important for the common user..."

The abundance of less significant or obscure topics can also contribute to the perceived sparsity. While Wikipedia's commitment to covering even niche subjects is commendable, it might be overwhelming for casual users seeking concise and essential information. For instance, niche historical events, minor fictional characters, or obscure scientific theories might exist as standalone articles, but their relevance to the everyday reader could be questioned. - ChatGPT

Source Data

Initial Data Collection and Normalization

[More Information Needed]

Who are the source language producers?

[More Information Needed]

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

[More Information Needed]

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

This article removes all "Notices" from all articles to provide a cleaner version of wikipedia.
You should consider adding flags back into the dataset if you want to tell the user about the potential issues.

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

KaraKaraWitch

Licensing Information

Most of Wikipedia's text and many of its images are co-licensed under the Creative Commons Attribution-ShareAlike 3.0 Unported License (CC BY-SA) and the GNU Free Documentation License (GFDL) (unversioned, with no invariant sections, front-cover texts, or back-cover texts).

Some text has been imported only under CC BY-SA and CC BY-SA-compatible license and cannot be reused under GFDL; such text will be identified on the page footer, in the page history, or on the discussion page of the article that utilizes the text.

Citation Information

@misc{superwiki,
  title         = {SuperWIKI Cleaned: Wikipedia for commoners.},
  author        = {KaraKaraWitch},
  year          = {2023},
  howpublished  = {\url{https://huggingface.co/datasets/RyokoExtra/SuperWIKI}},
}

Name Etymology

N/A

Contributions

@KaraKaraWitch (Twitter) for gathering this dataset.
@sirneggles (Twitter) for provided compute.