Prompt Engineering for Beginners: From Zero to Expert

Explore the foundations of AI, generative AI, LLM, and prompt engineering through a kitchen analogy, showing how prompts guide LLMs to create text, images, and novel content.

Explore how ChatGPT works atop a generative pre-trained transformer, its capabilities and limitations including LM hallucination, and how to interact with its interface—from login to settings and model options.

Explore how text is tokenized into tokens, embedded as vectors, and processed by a transformer with attention to generate the next token, guided by context windows and token limits.

Explore how ChatGPT handles logical reasoning questions by breaking down complex problems step by step, solving classic reasoning puzzles, and demonstrating clear, critical thinking.

Test deductive reasoning with questions that reveal how ChatGPT follows premises to reach conclusions, like ranking heights and tea or coffee preferences, and visualize results with mermaid diagrams.

Explore how prompts guide LLMs like ChatGPT to solve mathematical reasoning problems, using geometric progression, tables, and graphs to reveal patterns, prompt engineering, and worst-case scenarios.

This lecture tests llm pattern recognition with geometric progression tasks and code generation to extend series, showing how ChatGPT recognizes sequences, explains Fibonacci-based letter-number patterns, and discusses hallucination risks.

Examine common sense reasoning with ChatGPT, testing how LLMs understand human behavior, safety, and decision making via real-world scenarios such as buy-one-get-one offers and ethical dilemmas.

Explore temporal reasoning, time zone conversions, and duration calculations through practical questions, showing how LLMs handle sequences, time constraints, and context switches in real-world scenarios.

Explore how large language models simulate causal reasoning to map cause-and-effect relationships and predict outcomes. Analyze sleep's impact on memory consolidation and focus, and discuss AI advancement toward machine consciousness.

Explore how spatial reasoning works and how large language models approach space tasks by evaluating desk arrangements and floor plans, highlighting limits and practical design applications.