Quick Start Guide to Large Language Models (LLMs): ChatGPT, Llama, Embeddings, Fine-Tuning, and Multimodal AI, 2nd Edition

Introduction

Quick Guide to ChatGPT, Embeddings, and Other Large Language Models (LLMs): Introduction

Module 1: Introduction to Large Language Models

Module Introduction

Lesson 1: Overview of Large Language Models

Topics
1.1 What Are Large Language Models?
1.2 Popular Modern LLMs
1.3 Applications of LLMs

Lesson 2: Semantic Search with LLMs

Topics
2.1 Introduction to Semantic Search
2.2 Building a Semantic Search System
2.3 Optimizing Semantic Search with Cross-Encoders and Fine-Tuning

Lesson 3: First Steps with Prompt Engineering

Topics
3.1 Introduction to Prompt Engineering
3.2 Working with Prompts Across Models
3.3 Building a Retrieval-Augmented Generation BOT with ChatGPT and GPT-4

Lesson 4: Retrieval Augmented Generation + AI Agents

Topics
4.1 Introduction to Retrival Augmented Generation (RAG)
4.2 Building a RAG bot
4.3 Using Open Source Models with RAG
4.4 Expanding into AI Agents

Module 2: Getting the Most Out of LLMs

Module Introduction

Lesson 5: Optimizing LLMs with Fine-Tuning

Topics
5.1 Transfer Learning—A Primer
5.2 The OpenAI Fine-Tuning API
5.3 Case Study: Predicting with Android App Reviews—Part 1
5.4 Case Study: Predicting with Android App Reviews—Part 2

Lesson 6: Advanced Prompt Engineering

Topics
6.1 Input/Output Validation
6.2 Batch Prompting + Prompt Chaining
6.3 Chain-of-Thought Prompting
6.4 Preventing Prompt Injection Attacks
6.5 Assessing an LLM’s Encoded Knowledge Level

Lesson 7: Customizing Embeddings + Model Architectures

Topics
7.1 Case Study: Building an Anime Recommendation System
7.2 Using OpenAI’s Embedded Models
7.3 Fine-tuning an Embedding Model to Capture User Behavior

Lesson 8: AI Alignment–First Principles

Topics
8.1 Introduction to AI Alignment
8.2 Evaluating Alignment Plus Ethics

Module 3: Advanced LLM Usage

Lesson 9: Moving Beyond Foundation Models

Topics
9.1 The Vision Transformer
9.2 Using Cross Attention to Mix Data Modalities
9.3 Case Study—Visual QA: Setting Up Our Model
9.4 Case Study—Visual QA: Setting Up Our Parameters and Data
9.5 Introduction to Reinforcement Learning from Feedback
9.6 Aligning FLAN-T5 with Reinforcement Learning from Feedback

Lesson 10: Advanced Open-Source LLM Fine-Tuning

Topics
10.1 BERT for Multi-label Classification—Part 1
10.2 BERT for Multi-label Classification—Part 2
10.3 Writing LaTeX with GPT-2
10.4 Case Study: Sinan’s Attempt at Wise Yet Engaging Responses—Sawyer
10.5 Instruction Alignment of LLMs: Supervised Fine-Tuning
10.6 Instruction Alignment of LLMs: Reward Modeling
10.7 Instruction Alignment of LLMs: RLHF
10.8 Instruction Alignment of LLMs: Using Our Instruction-Aligned LLM

Lesson 11: Moving LLMs into Production

Topics
11.1 Cost Projecting and Deploying LLMs to Production
11.2 Knowledge Distillation

Lesson 12: LLM Evaluations

Topics
12.1 Evaluating Generative Tasks—Part 1
12.2 Evaluating Generative Tasks—Part 2
12.3 Evaluating Understanding Tasks
12.4 Probing LLMs for world model

Summary

Quick Guide to ChatGPT, Embeddings, and Other Large Language Models (LLMs): Summary