Concept

Notational Simplification in Fine-Tuning Formulas

For the sake of clarity and to keep mathematical notation uncluttered, the parameters being adjusted during fine-tuning, which are initialized from a pre-trained state θ^+\hat{\theta}^{+}, are often simplified in formulas to just θ\theta. Although the superscript '+' is omitted in this convention, it is crucial to remember that these parameters originate from the pre-trained model and are not randomly initialized.

0

1

Updated 2026-01-15

Contributors are:

Who are from:

Tags

Ch.4 Alignment - Foundations of Large Language Models

Foundations of Large Language Models

Computing Sciences

Foundations of Large Language Models Course