Analyzing Training Objectives in Model Adaptation
A large language model is first trained on a massive and diverse dataset from the public internet to develop a broad understanding of language patterns, facts, and reasoning. Subsequently, this same model undergoes a second phase of training using a much smaller, specialized dataset of legal documents to improve its performance on legal-related queries. Analyze the primary difference in the objective between the initial, broad training phase and the subsequent, specialized training phase.
0
1
Tags
Ch.2 Generative Models - Foundations of Large Language Models
Foundations of Large Language Models
Foundations of Large Language Models Course
Computing Sciences
Analysis in Bloom's Taxonomy
Cognitive Psychology
Psychology
Social Science
Empirical Science
Science
Related
Adapting a General LLM for a Specialized Task
A development team starts with a large, pre-trained language model that has a broad, general understanding of language. Their goal is to create a specialized tool that accurately classifies customer feedback into three specific categories: 'Positive', 'Negative', or 'Neutral'. They have a dataset of 50,000 customer feedback entries, each correctly labeled with one of the three categories. The team decides to use this labeled dataset to perform additional training on the model. Which statement best analyzes the primary purpose and mechanism of this adaptation process?
Analyzing Training Objectives in Model Adaptation