Enrolment options

Introduction to Large Language Models (LinkedIn Learning)

Course duration: 1,29h

Large language models (LLMs) have taken the AI world by storm. LLMs are behind some of the biggest AI technologies over the last few years, like ChatGPT and GPT-4. In this course, Jonathan Fernandes provides an overview of LLMs suitable for technical learners and non-technical learners alike. Jonathan shows you what LLMs are and what you can do with them, and takes a look under the hood so you can understand why they work the way they do and how they can affect your work. He explains how LLMs are trained and details the components of LLMs, and then takes a look at several different applications of LLMs—including Google’s BERT, GPT-3, PaLM and PaLM 2, ChatGPT and GPT-4, and Llama—and shows you how to compare LLMs using benchmarks.

Topics include:
  • Describe the architecture of large language models and their components such as transformers, encoders, and decoders.
  • Explain the function and importance of model parameters in neural networks.
  • Analyze the efficiency and effectiveness of different language models like GPT-3, BERT, and PaLM in various tasks.
  • Interpret the role of scaling laws in optimizing large language models' performance.
  • Evaluate the impact of model size and training data volume on the effectiveness of language models.
  • Identify the benefits and drawbacks of open community language models compared to proprietary models.
  • Compare language model performance using benchmarks like HELM and understand their implications.
  • Discuss the trends in reducing inference costs and improving model performance.
  • Illustrate the concept of tokenization and context windows in the generation of responses by language models.
  • Outline the key challenges and limitations of large language models, including their inability to update in real-time.

 

Guests cannot access this course. Please log in.