50 ML Projects To Understand LLMs: Investigate transformer mechanisms through data analysis, visualization, and experimentation

Mike X Cohen

This book teaches how LLMs like GPT and BERT actually work by applying machine learning techniques to their internal activations. It takes a unique approach: rather than building LLMs from scratch or using them via APIs, learners will investigate their mechanisms by treating hidden states, attention patterns, and embeddings as data to analyze.

Through 50 hands-on projects, you’ll learn to: 1) Inspect and visualize transformer internals; 2) Analyze attention mechanisms and layer dynamics; 3) Apply statistical and causal methods to understand model behavior; 4) Manipulate activations to test hypotheses about LLM mechanisms.

Each project teaches skills in three areas: machine learning techniques, LLM mechanisms, and Python coding with data visualization.