Parameters

Parameters are the billions of numerical values inside a model that were adjusted during training. They're not facts stored in a table — they're weights that collectively encode patterns across the entire model. When you hear "GPT-4 has hundreds of billions of parameters," that's the scale of this number grid. More parameters generally means more capacity, but size alone doesn't determine quality.

Videos explaining this concept

E005

Notes on AI

What Is a Model, Really?

A model is a learned structure that predicts what comes next. The training process — analyzing massive amounts of text data — doesn't stay "alive" inside the model. It collapses and crystallizes in...