Corpus
Domain: Data & Datasets
🎯 Motivation
A large collection of text used for training. Can include books, articles, websites, code.
📋 Concrete Example
Corpus: 5 million books, 20 million articles, 100 million web pages, 10 million code files. The model learns language patterns from this massive text collection.
🔗 Analogy
A library’s entire collection - millions of books on every topic. A corpus is the model’s library, the text it learns from.