Extrapolation

Appears in 1 paper

Extending a fitted line to predict values beyond the observed range.

As used in Paper 13 — Scaling Laws for Neural Language Models →

Extending a fitted line to predict values beyond the observed range. After fitting power laws to experiments with up to 100B parameters, you can extrapolate to 175B (GPT-3) or larger.

Paper 13 — Scaling Laws for Neural Language Models →