Module 2: Techniques in Prompting

prompting techniques or prompt patterns

  • Zero-shot prompting
  • Few-shot prompting
  • Chain-of-thought prompting
  • Instruction-style prompting
  • Role prompting or persona-based prompting

Each technique gives the model better instructions, depending on what we want.

Zero-shot Prompting

In zero-shot prompting, we just ask the question or give the instruction without giving any examples.

Example:

Prompt: “Translate to Hindi: ‘How are you?’”

Output: “Aap kaise hain?”

Even though we did not show any examples, the model understood the task. This is useful when:

  • You want a quick result

  • You believe the model already knows how to do the task.

Few-shot Prompting

In few-shot prompting, we give the model a few examples before asking it to do something new. This helps it understand the pattern better.

```

Example:

Prompt: English: Hello
Hindi: Namaste
English: Thank you
Hindi: Dhanyavaad
English: Good morning
Hindi:

Output: Shubh prabhat

This method is useful when:

  • You want better accuracy

  • You’re not sure if the model knows the task

  • You can give 2–3 examples

Chain-of-Thought Prompting (CoT)

This is a special technique where we ask the model to think step-by-step.

Example:

Prompt:

I went to the market and bought 10 apples. I gave 2 apples to the neighbor and 2 to the repairman. I then went and bought 5 more apples and ate 1. How many apples did I remain with? Let’s think step by step.

Output:

First, you started with 10 apples. You gave away 2 apples to the neighbor and 2 to the repairman, so you had 6 apples left. Then you bought 5 more apples, so now you had 11 apples. Finally, you ate 1 apple, so you would remain with 10 apples.

Chain-of-thought helps in:

  • Math problems

  • Logical reasoning

  • Multi-step tasks

Instruction-style Prompting

Here, we give clear instructions in a natural language format. This is how ChatGPT and InstructGPT were trained.

Example:

Prompt:

“Write a short summary of the movie Inception in 2-3 lines using simple English.”

Output:

Inception is a sci-fi movie where people enter dreams. A team tries to plant an idea in someone’s mind through dreams.

Instruction-style prompts are useful for:

  • Summarization

  • Content creation

  • Rewriting tasks

  • Questions and answers

Role or Persona Prompting

In this type, we ask the model to act like someone. This helps set a context or tone.

Example:

Prompt:

“You are a nutritionist. Explain why breakfast is important to a 10-year-old.”

Output:

Breakfast gives your body energy after sleeping all night. It helps you focus in school and keeps you strong and healthy.