Assessment: Transformers and LLMs

Explain

• What problem does attention solve?
• Why do transformers use positional information?
• What is the difference between pretraining and fine-tuning?
• What does temperature control?
• Why do hallucinations happen?

Build

• Design a structured output workflow.
• Design a function calling workflow.
• Explain how you would compare two models.

Debug

• The model returns malformed JSON. What do you do?
• The model is too expensive. What options do you evaluate?
• The model ignores instructions. What do you test?

Completion Gate

• Can implement attention basics.
• Can integrate an LLM safely.
• Can validate structured outputs.