Training data
The large collection of text an AI model learned from before it was released.
In plain English
What a model “knows” without searching comes from its training data, which stops at a cutoff date. If your business was poorly described online when the model was trained, the model may know little about you until it searches live.
In practice
A café that opened last year is unknown to a model trained before then, unless the AI looks it up on the live web.
Training datavsGrounding
Training data is what the model learned in the past. Grounding is what it looks up live before answering.
Related terms
How AI reads the webStarter Knowledge cutoff The date after which an AI model has no built-in knowledge. In practice, and related terms → How AI reads the webPro Grounding When an AI checks live sources before answering, instead of relying only on what it learned in training. In practice, and related terms → How AI reads the webStarter Large language modelLLM The kind of AI behind ChatGPT, Claude and Gemini, trained on huge amounts of text to understand and write language. In practice, and related terms →