Fine-Tuning vs Prompting Trade-offs Quiz

What you'll be asked

1. Which of the following scenarios is most appropriate for using few-shot prompting instead of fine-tuning?
2. What is the primary cost-saving advantage of fine-tuning a smaller base model compared to using a larger base model with long-context prompting?
3. When evaluating the cost-effectiveness of fine-tuning, which factor should be considered as a 'hidden' cost?
4. In a RAG (Retrieval-Augmented Generation) architecture, when would fine-tuning be most beneficial for cost optimization?
5. An enterprise application requires 99.9% accuracy on a specialized classification task. A base model with few-shot prompting yields 85% accuracy. If fine-tuning improves accuracy to 98% but requires an expensive retraining cycle every time the data distribution shifts, which statement best reflects the cost-optimization trade-off?

Answers and explanations come after you submit.