LLM
Articles on large language models — attention, training, inference, evaluation, and the systems around them. Standalone pieces rather than a strict curriculum.
About This Series
Each article tackles a specific question or mental model around LLMs — what attention is really doing, how training loops actually run, why inference behaves the way it does, and the practical engineering that surrounds production deployments. Read in any order.
Articles
Attention resolves through DNS
Transformer attention from first principles using the DNS mnemonic — Dot, Normalize, Sum — with interactive visualizations.
• 26 min read
More on the way
New articles are added as topics come up. Follow the blog or check back for the next piece.