III. Lo Scrittoio

ArticlesNotes from the scriptorium

  1. How language models workAn explanation without the math: how a language model guesses its way forward one word at a time, what attention really is, and why it can sound convincing even when it is wrong.Foundations12 min
  2. How Claude Skills workSkills are self-contained instruction sets Claude loads on demand. What they’re made of, how they load in layers, and why they turn Claude into a specialist in your field.Foundations5 min
  3. How tokens work in language modelsWhy can’t ChatGPT count the letters in “strawberry”? The answer is tokens: the pieces of text the model actually reads, never the letters themselves.Foundations11 min
  4. Hyperparameters in the prompting processTemperature and Top P are usually explained as “randomness”. What they actually do to the words a model chooses between, and why temperature 0 doesn’t always give the best answers.Foundations9 min
  5. Prompt engineering: It’s not about instructions, it’s about contextWhy language models don’t follow instructions the way a computer follows code, and why good prompting is about building the right context.Foundations8 min