Fine-Tuning in Practice

1. What does fine-tuning primarily aim to do with a pre-trained model?

Increase its parameter count Change its foundational architecture Adjust its behavior to better match specific guidelines Inject new factual knowledge

2. Which technique is highlighted as a method for parameter-efficient fine-tuning?

Meta Learning LoRA (Low-Rank Adaptation) Gradient Descent Transfer Learning

3. What is fine-tuning NOT effective at achieving?

Consistent output format Domain vocabulary mastery Injecting new knowledge Model style and tone adaptation

4. Why is it important to evaluate before and after fine-tuning a model?

To increase the dataset size To confirm that fine-tuning improves model performance To change the architecture To add more training data

Continue: Choosing a Strategy → Review this lesson Retake quiz