How LLMs Work (Enough to Architect With)

1. What is the primary function of a large language model (LLM)?

To retrieve facts from a database To predict the next token in a sequence To summarize large documents To replace human writers

2. Why is it important for architects to understand that models read and write tokens instead of words?

Tokens are easier to process than full words Costs and limits in AI models are measured in tokens Tokens allow for more complex sentences Models only understand technical jargon

3. What does the term 'context window' refer to in large language models?

The total amount of data the model can train on The fixed amount of tokens that can be processed in a single call The variables used to train the model The size of the model's vocabulary

4. What role does 'context' play in the output of a large language model?

It defines the model's historical data It serves as the primary mechanism for fresh, specific information at call time It limits the creativity of the model It is the model's core algorithm

Continue: Capabilities and Limits → Review this lesson Retake quiz