Neural scaling laws
Scaling-law research models loss as power-law relationships with compute, data, and parameter count over studied regimes. An empirical scaling trend supports planning within its evidence range but is not a physical law guaranteeing every architecture or dataset.
Overview. Scaling-law research models loss as power-law relationships with compute, data, and parameter count over studied regimes.
Context. An empirical scaling trend supports planning within its evidence range but is not a physical law guaranteeing every architecture or dataset.
Claims and evidence
Claim ledger, 2Scaling-law research models loss as power-law relationships with compute, data, and parameter count over studied regimes.
An empirical scaling trend supports planning within its evidence range but is not a physical law guaranteeing every architecture or dataset.