LLM Training Data

GEO & AEO

LLM training data is the massive set of texts and information used to train a language model (Large Language Model), from which the model learns linguistic patterns, facts, and associations, often including public web content, books, articles, and other text sources.

Content from a website can end up in future models’ training data if it is accessible to AI-type crawlers (such as GPTBot or CCBot) and the site does not explicitly block these crawlers via robots.txt.

Why it matters

The decision to allow or block AI crawlers from collecting training data is an increasingly relevant strategic decision for brands: some want to be represented in future models, others prefer to protect their original content from unauthorized reuse.

← All terms in GEO & AEO Full glossary