Engineering

Examples of Prompts 📝

Explore various real-world examples of how to use prompts for tasks like summarization, extraction, classification, and code generation.

Driptanil DattaSoftware Developer

Mar 20268 min read

Learning through Examples

The best way to master prompt engineering is by exploring how different prompts lead to different model behaviors. This section provides concrete examples of common tasks and how to structure your prompts for them.

1. Text Summarization

LLMs are exceptionally good at distilling long paragraphs into concise summaries.

Antibiotics are a type of medication used to treat bacterial infections... [Long Paragraph] ... 
 
Explain the above in one sentence:

Antibiotics are medications used to treat bacterial infections by either killing the bacteria or stopping them from reproducing, but they are not effective against viruses and overuse can lead to antibiotic resistance.

2. Information Extraction

You can use prompts to pull specific data points out of unstructured text.

[Research Paper Paragraph] ... Mention the large language model based product mentioned in the paragraph above:

The large language model based product mentioned in the paragraph above is ChatGPT.

3. Question Answering

By providing context and specific instructions, you can improve the factual accuracy of the model's responses.

Answer the question based on the context below. Keep the answer short and concise. Respond "Unsure about answer" if not sure.
 
Context: [Technical details about Teplizumab]
Question: What was OKT3 originally sourced from?
Answer:

Mice.

4. Text Classification

You can guide the model to use specific labels and formats by providing examples (few-shot prompting).

Classify the text into neutral, negative or positive. 
 
Text: I think the vacation is okay. 
Sentiment: neutral 
 
Text: I think the food was okay. 
Sentiment:

neutral

5. Conversation & Role Prompting

One of the most powerful techniques is telling the LLM how to behave by assigning it a role.

6. Code Generation

LLMs can generate everything from simple scripts to complex database queries.

MySQL Query Generation:

"""
Table departments, columns = [DepartmentId, DepartmentName]
Table students, columns = [DepartmentId, StudentId, StudentName]
 
Create a MySQL query for all students in the Computer Science Department
"""

Output:

SELECT StudentId, StudentName 
FROM students 
WHERE DepartmentId IN (
    SELECT DepartmentId 
    FROM departments 
    WHERE DepartmentName = 'Computer Science'
);

Conclusion

As you've seen, LLMs can perform a vast range of tasks simply through clever instruction. In the next guide, we will dive into Prompting Techniques like Zero-Shot and Few-Shot prompting to further enhance these capabilities.