LLM Foundations

How large language models work, and how to build on them — from tokenization through retrieval, adaptation and evaluation.

For. Engineers comfortable with code and notation, working up from how a model processes text to how one is adapted and evaluated.

  1. 1
    How Large Language Models Actually Work

    ~15 min

    Four mechanisms, from raw text to generated token

  2. 2
    Embeddings and Retrieval

    ~12 min · follows How Large Language Models Actually Work

    How a model answers questions about things it never saw in training