Scaling Hypothesis / Scaling Laws

Definition

Empirical model showing that large language model capabilities improve systematically with increases in compute, data, and model parameters across training runs. The relationship follows a power law, allowing prediction of future capabilities by training at larger scales. Not based on fundamental physical law but on consistent empirical observation.

Cast within

No topic-region cast yet — this appears once Scaling Hypothesis / Scaling Laws's compiled claims are aligned into a topic region's argument tree.

Claims mentioning Scaling Hypothesis / Scaling Laws (9)

Your notes

Loading notes...