How models work · L2 · 15 min
Model cards and ids
Learn in five minutes
- A model id is a name you pin, because names move.
- A model card should say limits, data, and intended use. Confirm the card you rely on.
- Open-weight and hosted models are different operational choices.
- This lab does not rank this month's products.
- Write 'confirm before relying' next to any id you did not re-check.
Copying a model id from a 2024 blog into a 2026 script without checking is how labs rot. Pin the id in your own note and date the check.
Visual
A pin line
- Write the provider, the id, and the date you looked.
- Add the sentence 'confirm before relying' if you did not look today.
- Do not add a price.
Quiz
Ecosystem
- Model card
- The provider's note of intended use and limits. Confirm it is current.
Sisters: Tokenizers, Transformers, Decoding, Model cards.
The same next-token idea sits under chat, voice, and tool calls.
Pin a model id when you leave the toy explanations.
Config
- Write the goal in one sentence.
- Name the input and the allowed output.
- Save the trace next to the result.
Troubleshooting
- The answer looks fluent and is wrong. Check whether a source passage actually contains the claim.
- The loop never stops. Set a step cap and stop when the tool result repeats.
Sources and go deeper
- NIST AI Risk Management Framework spec last checked 2026-10
- Brown and co-authors, Language Models are Few-Shot Learners, 2020 paper last checked 2026-10
- Large language model — Wikipedia wikipedia last checked 2026-10