Few-shot prompting in Prompt Engineering / GenAI - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding Few-shot Prompting Basics

What is the main advantage of few-shot prompting when using large language models?

A It provides a few examples in the prompt to guide the model's output.

B It allows the model to learn new tasks without any examples.

C It requires retraining the model with a large dataset.

D It reduces the model size to improve speed.

💡 Hint

Think about how few-shot prompting helps the model understand what you want by showing examples.

Few-shot prompting works by giving the model a few examples in the prompt so it can mimic the pattern and produce the desired output.

❓ Predict Output

intermediate

Predicting Output from Few-shot Prompt

Given this prompt to a language model, what is the most likely output? Prompt: "Translate English to French: 1. Hello -> Bonjour 2. Thank you -> Merci 3. Good night ->"

A Bonne nuit

B Bonsoir

C Salut

D Au revoir

💡 Hint

Look at the pattern of English phrases and their French translations.

The prompt shows English phrases with their French translations. 'Good night' translates to 'Bonne nuit' in French.

❓ Model Choice

advanced

Choosing the Best Model for Few-shot Prompting

Which type of model is best suited for few-shot prompting tasks?

A Small models trained only on specific tasks

B Large pretrained language models with broad knowledge

C Models trained only on numerical data

D Models without any pretraining

💡 Hint

Few-shot prompting relies on the model's prior knowledge and ability to generalize.

Large pretrained language models have learned broad patterns and knowledge, making them effective for few-shot prompting.

❓ Hyperparameter

advanced

Effect of Temperature in Few-shot Prompting

In few-shot prompting, what effect does increasing the temperature parameter have on the model's output?

A Decreases the length of the output

B Makes the output more deterministic and repetitive

C Makes the output more random and creative

D Stops the model from generating any output

💡 Hint

Temperature controls randomness in the model's choices.

Higher temperature values increase randomness, making outputs more diverse and creative.

❓ Metrics

expert

Evaluating Few-shot Prompting Performance

Which metric is most appropriate to evaluate the quality of few-shot prompting outputs for a text classification task?

A BLEU score

B Mean Squared Error (MSE)

C Perplexity

D Accuracy

💡 Hint

Think about the task type and what metric measures correct predictions.

Accuracy measures how many predictions match the true labels, suitable for classification tasks.