GenAI in production · pillar 1 of 6
How the thing behaves
Tokens, context and sampling — the three mechanics every surprise traces back to.
- 1 It reads tokens, not words or characters A model does not see text. It sees tokens — sub-word fragments produced by a tokeniser specific to that model family. A…
- 2 The context window is not memory Each request is stateless. What looks like memory is the entire conversation being re-sent every turn, inside the…
- 3 The output is sampled, not looked up A model produces a probability distribution over the next token and then one is sampled from it. Settings such as…
Start this pillar See the whole map
Read against OWASP Top 10 for Large Language Model Applications, 2025, NIST AI 600-1 — Generative AI Profile (AI RMF), Lewis et al., Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks (arXiv:2005.11401), Hu et al., LoRA: Low-Rank Adaptation of Large Language Models (arXiv:2106.09685) on . Regulations and frameworks change; the date is there so you can tell whether this has kept up.
The OWASP Top 10 for LLM Applications is © the OWASP Foundation and published under a Creative Commons licence; NIST AI 600-1 is a US Government work in the public domain. These cards are independent commentary written in our own words. ytyk.org is not affiliated with or endorsed by the OWASP Foundation or NIST.