How the Machine Actually Works

1. What is the main task that a large language model (LLM) is trained to perform?

Translate languages Summarize articles Predict the next word Generate images

2. What term describes the model's knowledge that is stored and frozen at the moment training ends?

Parametric knowledge Dynamic knowledge Retrieval knowledge Contextual knowledge

3. What do the weights in an LLM represent?

The documentation of the model The current input text A compressed model of the world The tools available to the agent

4. What role does the context window play during the model's answer generation process?

It stores the complete training data It helps in generating new training data It provides current and citable information It eliminates hallucinations completely

Continue: How LLMs Work (Enough to Architect With) → Review this lesson Retake quiz