Situating LLMs in Time: A Visual Tool for Teaching the History and Architecture of Modern NLP
Understanding the history of technology is essential to teaching computer science, not just how systems work, but how we got there. This is more true in the rapid evolution of Large Language Models (LLMs), whose development since 2017 offers a rich, branching narrative of innovation and adaptation. To support history-aware computing education, we introduce an interactive visual timeline that traces the lineage of LLMs, and different techniques within, from the foundational “Attention is All You Need” paper through major model families and research directions. Our tool, LLM Timeline, is organized as a top-down graph with color coded branches that represent various architectures and approaches such as Decoder-Only, Encoder-Only, Mixture-of-Experts, Multimodal, and much more, allowing students to explore not only technical contributions but also the intellectual lineage of modern AI.