RWFork: Russian Wikipedia Fork Edits Dataset

Description

This dataset contains the edits made by editors of the Russian Wikipedia Fork (RWFork, ruwiki.ru) to the original Russian Wikipedia articles between May and September 2023. It includes detailed textual differences between the two versions of the encyclopedia, metadata changes, and additional information such as edit summaries and embeddings. These features provide a comprehensive overview of the editorial activity and enable further analysis and processing.

Keywords: RWFork, Wikipedia, content differences, knowledge manipulation, data mining, text analysis, computational social science, propaganda.

Dataset Details

Language: Russian

Size:

33,664 rows, 17 columns (main dataset)

17,745 rows, 1 column (pages that we were not able to find, extract, or parse from RWFork)

Format: CSV

License: CC BY 4.0

Use Case: Text mining, knowledge manipulation detection and analysis, content moderation.

Processing details: The dataset was created by comparing the wikitext of the Russian Wikipedia and RWFork articles with the same title, extracting fine-grained edits: additions, deletions, and changes to the page content (text) and metadata. Only pages with content differences were included, excluding articles that were identical between the two versions or not found in RWFork. Additional features (e.g., edit summary, embedding) were calculated using OpenAI API.

Files: rwfork_changed.csv (main dataset), rwfork_columns.md (descriptions of main dataset columns), rwfork_not_found.csv (pages names that we were not able to find, extract, or parse from RWFork).

Attribution

The dataset was compiled from Russian Wikipedia articles and its corresponding fork.
Below, we provide information about the original content's license and the template to generate the link to the original source for a given page (). For example, https://ru.wikipedia.org/wiki/Москва_(ракетный_крейсер) links to the page “Москва_(ракетный_крейсер)” in Russian Wikipedia.

Wikipedia

Source: https://ru.wikipedia.org/wiki/

License: CC BY-SA 4.0, GFDL

RWFork

Source: https://ru.ruwiki.ru/wiki/

License: CC BY-SA 4.0, GFDL

Limitations: The dataset is created from community-edited content and may contain inaccuracies, vandalism, or other forms of knowledge manipulation in both versions.

Potential Use Cases:

Analysis and detection of russian knowledge manipulation strategies in encyclopedic content.

Development of algorithms for content moderation and vandalism detection.

Understanding the dynamics of forked communities and their impact on the original content.

Studying the differences in editorial practices and content quality between Wikipedia and its forks.

Training and evaluation language models for downstream tasks using fine-grained edit data.

Please cite as:

TBD