Your question is Fine-Tuning vs In-Context Learning. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
What are the primary trade-offs when choosing between fine-tuning a smaller model versus using a larger model with in-context learning?
Explain how you would make the decision for a production LLM feature. Address quality, latency, inference and training cost, data requirements, adaptation speed, context limits, maintainability, privacy, hallucination risk, and prompt injection. Describe an evaluation plan that compares both approaches on the same representative workload, including offline and online metrics. State when you would choose each option and whether a hybrid or routing strategy is appropriate.