Editorial equation guide
Pretraining compute
The one-time computation used to learn from the initial training data.
Symbol in context
This entry links every published equation using this exact notation. Its meaning may change between equations.
Editorial equation guide
The one-time computation used to learn from the initial training data.
Article-grounded definition
pretraining compute.
This equation states an equality: the expressions on both sides have the same value under the article’s assumptions.
Equation guide → · This term → · Article →This equation gives an approximation: it relates the quantities while allowing an approximation.
Equation guide → · This term → · Article →The total compute bill has three parts: pretraining, post-training, and all the requests served afterwards.
Equation guide → · This term → · Article →This equation gives an approximation: it relates the quantities while allowing an approximation.
Equation guide → · This term → · Article →