# Dataset Card for Enwiki Dataset

This is an automatically updating dataset containing ~7 million English Wikipedia articles, with expanded templates and converted into Markdown. This dataset was created with the intention to provide a bite-sized, LLM-readable version of Wikipedia for various applications, including RAG.

## Dataset Overview

There are two main versions of the dataset:

- **`merged-articles`** - Complete Wikipedia dump with all articles merged into a single file.
- **`merged-article-chunked`** - Articles chunked into ~700 word segments with Markdown-header hierarchical breadcrumbs (Gemini embeddings coming soon!)

The latest version of the dataset was updated on `2025-08-16`.

Blogpost coming soon!