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...