When to fine-tune vs prompt engineer in Prompt Engineering / GenAI - Quick Revision & Key Differences

Recall & Review

beginner

What is fine-tuning in the context of AI models?

Fine-tuning means adjusting a pre-trained AI model by training it a bit more on a specific task or data to make it better at that task.

beginner

What does prompt engineering involve?

Prompt engineering is about carefully writing or designing the input text (prompt) to guide an AI model to give the best possible answer without changing the model itself.

intermediate

When should you choose fine-tuning over prompt engineering?

Choose fine-tuning when you need the model to deeply understand a special task or data that is very different from what it learned before, or when prompt engineering can't get good enough results.

intermediate

What are the benefits of prompt engineering compared to fine-tuning?

Prompt engineering is faster, cheaper, and doesn't need extra training. It works well when you want quick changes or when the task is similar to what the model already knows.

intermediate

Give an example scenario where fine-tuning is preferred.

If you want an AI to understand medical reports in a special format that it hasn’t seen before, fine-tuning with medical data helps the model learn those details better than just changing the prompt.

Which method involves changing the input text to guide the AI model's output?

A Prompt engineering

B Fine-tuning

C Model architecture redesign

D Data cleaning

Prompt engineering means crafting the input prompt to get better answers without changing the model.

When is fine-tuning usually necessary?

A When you have no data for training

B When you want to quickly test different prompts

C When the task is very different from the model's original training

D When you want to reduce model size

Fine-tuning helps the model learn new, specific tasks that differ from its original training.

What is a key advantage of prompt engineering over fine-tuning?

A It changes the model weights

B It requires no extra training

C It needs large datasets

D It takes longer to implement

Prompt engineering works by changing inputs, so no extra training is needed.

Which approach is better for adapting a model to a very specialized domain with unique data?

A Fine-tuning

B Prompt engineering

C Using default prompts

D Random guessing

Fine-tuning adjusts the model to understand specialized data better.

If you want to quickly improve AI responses without retraining, what should you do?

A Collect more data

B Fine-tune the model

C Change the model architecture

D Use prompt engineering

Prompt engineering allows quick improvements by changing the input prompts.

Explain the main differences between fine-tuning and prompt engineering.

Think about what changes: the model or the input?

Describe a situation where fine-tuning is necessary and why prompt engineering would not be enough.

Consider tasks very different from the model's original training.