Learn Before
Case Study

Choosing a Fine-Tuning Strategy

A startup is considering a tuning method where a large, pre-trained model's original parameters are kept frozen. For each new customer task, a small set of new, trainable vectors is created and prepended to the hidden states at each layer of the model to guide its behavior. Based on the startup's constraints described in the case study, evaluate the suitability of this proposed method and justify your conclusion.

0

1

Updated 2025-10-02

Contributors are:

Who are from:

Tags

Ch.3 Prompting - Foundations of Large Language Models

Foundations of Large Language Models

Foundations of Large Language Models Course

Computing Sciences

Evaluation in Bloom's Taxonomy

Cognitive Psychology

Psychology

Social Science

Empirical Science

Science