×
\
Bonish Agarwal
Software Engineer
Bonish Agarwal is a Software Engineer at Nitor Infotech. He brings a wealth of expertise in Machine Learning and Generative AI to the table. ... Read More

In the fast-paced landscape of technological advancement, our interaction with machines is evolving at an unprecedented rate. Technologies like Artificial Intelligence (AI) are continuously broadening their understanding and harnessing extensive datasets to boost machine learning. Thus, providing a plethora of golden opportunities for businesses.

When talking about AI, let’s not overlook the buzz of Generative AI and prompt engineering in this modern tech scenario.

Let me give better clarity on this!

Picture this: You’re racing against the clock to finish a project but short on manpower. Panic sets in and that’s when prompts come to the rescue!

Prompts are like conversing with a highly intelligent friend who is a machine or an AI Model.

To accomplish your tasks, simply toss a thoughtfully crafted question or instruction—a prompt. This prompt functions as a magical bridge, enabling the machine to grasp your needs and respond with human-like understanding.

Voila! Your project will be saved, and delivered on time, and you will have spared yourself the headache of extra effort and expense.

Learn how we leveraged GenAI to categorize patients and tailor personalized care plans.

Interesting, right?

Well, further let’s get to know the world of prompt engineering, its techniques, applications, and the future it holds — all within the pages of this blog.

Happy reading!

What is Prompt Engineering?

At its core, prompt engineering is a nuanced art form. It’s about creating inputs that help Large Language Models (LLMs) not just understand but also process and give the responses you want. Think of it as a middleman, making sure what you want from the machine matches up with what it gives back, creating a smooth interaction.

The concept of prompt engineering may resonate with anyone who has interacted with voice assistants such as Amazon Alexa and Siri.

The way we formulate and structure requests to these assistants, such as saying, ‘Play songs by Taylor Swift’ versus ‘Play English music,’ can significantly influence the resulting responses you receive.

Today, prompting is a big deal in shaping how we interact. Think about systems like ChatGPT—your choice of words can seriously affect the responses you get, making a noticeable difference in the nature and quality of the conversation.

So, what’s the key to crafting a well-structured prompt that effectively communicates your intentions to the LLM?

The answer lies around knowing several prompting techniques that can guide you in composing prompts that are both clear and succinct, ensuring maximum results for your business.

Let’s check them out!

Prompt Engineering Techniques

Prompt Engineering Techniques

Here are some of the top techniques that will help you master the fundamentals of prompt engineering and scale your business in the GenAI landscape:

1. Zero-Shot Prompting: In this approach, the model autonomously comprehends the given prompt using its pre-existing knowledge base.

Refer to this example to understand it better:

Prompt:

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

Output:

Neutral

2. Few Shot Prompting: This technique involves presenting the model with a few examples within the prompt to enhance context comprehension. Thus, leading to improved performance.

Example:

Prompt:

Just answer according to the example given
India:India has a total of 28 states and the population is 1.4 billion.
Brazil:Brazil has a total of 26 states and the population is 0.22 billion.
America:

Output:

America: America is a country comprised of 50 states, and the population 
is approximately 331 million.

3. Chain of Thought Prompting: This technique involves guiding the model through a series of steps. It is used to break down a complex task into a chain of smaller steps for a better understanding of the model. It is the same as guiding a person step by step to solve an intricate math problem. This technique is primarily used to solve complex reasoning and math problems.

Example:

Prompt:

The odd numbers in this group add up to an 
even number: 4, 8, 9, 15, 12, 2, 1.
A: Adding all the odd numbers (9, 15, 1) gives 25. The answer is False.

The odd numbers in this group add up to an 
even number: 17, 10, 19, 4, 8, 12, 24.
A: Adding all the odd numbers (17, 19) gives 36. The answer is True.

The odd numbers in this group add up to an 
even number: 16, 11, 14, 4, 8, 13, 24.
A: Adding all the odd numbers (11, 13) gives 24. The answer is True.

The odd numbers in this group add up to an 
even number: 17, 9, 10, 12, 13, 4, 2.
A: Adding all the odd numbers (17, 9, 13) gives 39. The answer is False.

The odd numbers in this group add up to an 
even number: 15, 32, 5, 13, 82, 7, 1.
A:

Output:

Adding all the odd numbers (15, 5, 13, 7, 1) gives 41. 
The answer is False.

4. Self-Consistency: This advanced prompting technique enhances the chain of thought approach, particularly beneficial for tackling intricate reasoning problems. It entails sampling from multiple reasoning paths using a few-shot chain of thought prompting and subsequently utilizing the generated responses to select the most consistent and optimal answer.

Example:

Prompt:

Q: Shawn has five toys. For Christmas, he got two toys each from his 
mom and dad. How many toys does he have now?
A: He has 5 toys. He got 2 from mom, so after that he has 5 + 2 = 7 toys.
Then he got 2 more from dad, so in total he has 7 + 2 = 9 toys. 
The answer is 9.

Q: There were nine computers in the server room. Five more computers 
were installed each day, from monday to thursday. How many computers are 
now in the server room?
A: There are 4 days from monday to thursday. 5 computers were added 
each day.That means in total 4 * 5 = 20 computers were added. 
There were 9 computers in the beginning, so now there are 
9 + 20 = 29 computers.
The answer is 29.

Q: Michael had 58 golf balls. On tuesday, he lost 23 golf balls. 
On wednesday, he lost 2 more. How many golf balls did he have at the 
end of wednesday?
A: Michael initially had 58 balls. He lost 23 on Tuesday, so after that 
he has 58 - 23 = 35 balls. On Wednesday he lost 2 more so now he has 
35 - 2 = 33 balls.
The answer is 33.

Q: Olivia has $23. She bought five bagels for $3 each. 
How much money does she have left?
A: She bought 5 bagels for $3 each. This means she spent $15. 
She has $8 left.

Q: When I was 6 my sister was half my age. 
   Now I’m 70 how old is my sister?
A:

Output:

When you were 6 years old, your sister was half your age, which 
means she was 6 / 2 = 3 years younger than you.
Now that you are 70 years old, your sister’s age would be 
70 – 3 = 67 years old.
So, your sister is 67 years old now.

5. General Knowledge Prompting: This intriguing technique involves enriching your prompt by incorporating supplementary knowledge before making the OpenAI API call. Sometimes, to provide the necessary context and enhance the model’s understanding of the query, it’s essential to integrate additional knowledge. To address this, we can prompt the model to generate supplementary knowledge, which then can be integrated with our initial prompt to produce the desired outputs.

Here’s an example:

1st Prompt:

Input: Greece is larger than mexico.
Knowledge: Greece is approximately 131,957 sq km, while Mexico is 
approximately 1,964,375 sq km, making 
Mexico 1,389% larger than Greece.

Input: Glasses always fog up.
Knowledge: Condensation occurs on eyeglass lenses when water vapor from 
your sweat, breath, and ambient humidity lands on a cold surface, cools, 
and then changes into tiny drops of liquid, forming a film that you see 
as fog. Your lenses will be relatively cool compared to 
your breath, especially when the outside air is cold.

Input: A fish is capable of thinking.
Knowledge: Fish are more intelligent than they appear. In many areas, 
such as memory,their cognitive powers match or exceed those of ’higher’ 
vertebrates including non-human primates. Fish’s long-term memories help 
them keep track of complex social relationships.

Input: A common effect of smoking lots of cigarettes in one’s lifetime 
is a higher-than-normal chance of getting lung cancer.
Knowledge: Those who consistently averaged less than one cigarette per 
day over their lifetime had nine times the risk of dying from lung cancer 
than never smokers. Among people who smoked between one and 10 cigarettes 
per day, the risk of dying from lung cancer was nearly 12 times 
higher than that of never smokers.

Input: A rock is the same size as a pebble.
Knowledge: A pebble is a clast of rock with a particle size of 4 to 
64 millimeters based on the Udden-Wentworth scale of sedimentology. 
Pebbles are generally considered larger than granules (2 to 4 
millimeters diameter) and smaller than cobbles (64 to 256 millimeters 
diameter).

Input: Part of golf is trying to get a higher point total than others.
Knowledge:

Output:

In golf, the objective is to complete each hole in as few strokes 
as possible. Unlike some sports where a higher score indicates better 
performance, in golf, a lower score is better. The total number of 
strokes a player takes over the entire course determines their score. 
Each hole on a golf course has a designated "par" score, representing 
the number of strokes an expert golfer should take to complete it. 
Players aim to finish each hole in fewer strokes than the par score. 
The player with the lowest total score at the end of the round or 
tournament is the winner. So, in golf, the goal is to achieve 
a lower point total than others, not a higher one.

2nd Prompt:

Question: Part of golf is trying to get a higher point total than others. 
Yes or No?
Knowledge: In golf, the objective is to complete each hole in as few 
strokes as possible. Unlike some sports where a higher score indicates 
better performance, in golf, a lower score is better. The total number 
of strokes a player takes over the entire course determines their score. 
Each hole on a golf course has a designated "par" score, representing the 
number of strokes an expert golfer should take to complete it. 
Players aim to finish each hole in fewer strokes than the par score. 
The player with the lowest total score at the end of the round or 
tournament is the winner. So, in golf, the goal is to achieve a 
lower point total than others, not a higher one.
Explain and Answer:

Final Output:

No, part of golf is not trying to get a higher point total than others. 
In golf, the objective is to complete each hole in as few strokes 
as possible. A lower score is considered better in golf, and the total 
number of strokes a player takes over the entire course determines their 
score. Players aim to finish each hole in fewer strokes than the 
designated "par" score. The player with the lowest total score at the 
end of the round or tournament is the winner. Therefore, the goal in 
golf is to achieve a lower point total than others, not a higher one.

Curious about where you can harness such a remarkable ability? The following section delves into its various applications.

Keep reading!

Applications of Prompt Engineering

Applications of prompt engineering are vast and varied, like enhancing search engine performance, improving the quality of AI-generated creative writing, and fine-tuning responses for customer service chatbots. It also includes educational tools, wherein it can personalize learning experiences, and in programming, wherein it can support code generation and debugging. Here are the top applications of prompt engineering:

  • Content Generation
  • Generating code
  • Summarization

Here are some of the use cases of prompt engineering along with their examples:

1. Content Generation: Using prompt engineering is a common method in content creation. Whether you’re a seasoned content writer or an active social media creator, you can streamline the process and deliver the desired results. With a clear prompt, your next lead-gen email is just a step away.

I attempted to provide a prompt to ChatGPT, and you can see the response below –

Content Generation Response

Image generation with prompt engineering is a realm that combines human input’s creativity with artificial intelligence’s computational power. Via this process, users can provide detailed text prompts to AI models. They then interpret these instructions to create vivid, often stunning visual content.

This technology leverages deep learning algorithms, particularly Generative Adversarial Networks (GANs), to yield images that range from realistic photographs to fantastical art pieces.

The key to successful image generation lies in the art of prompt engineering — crafting prompts that are specific, evocative, and aligned with the desired outcome.

As AI continues to evolve, prompt engineering is becoming an essential skill for artists and creators, enabling them to harness the full potential of AI-assisted artistry.

2. Generating Code: If you’re a developer seeking a knowledgeable coding companion to assist you with sample codes and references, prompt engineering is your solution. By effectively applying the techniques mentioned above, you can obtain the precise code you need for your project.

Here’s what I tried generating –

Code Generation Response

3. Summarization: Summarizing extensive documents can undoubtedly be a time-consuming task. However, with the aid of prompt engineering, you can now summarize any document efficiently within a remarkably short timeframe. Thus, helping you manage complex tasks seamlessly.

Here’s what I tried summarizing with ChatGPT:

Summarization Response

Explore further insights into the applications of prompt engineering.

QUICK NOTE: The profound potential of cutting-edge AI technologies, when employed for positive purposes, can usher in transformative changes across diverse industries. However, this power also harbors the risk of misuse and harm.

Get the whole gist further as you feed your brain cells.

Vulnerabilities in Prompt Engineering

Here are the various types of vulnerabilities involved:

  • Prompt Injection: this means cleverly changing how LLMs behave by using well-thought-out prompts.
  • Prompt Leaking: a type of prompt injection when prompts are used to intentionally reveal information, like confidential details, not meant for public disclosure.
  • Jailbreaking: This is a smart way of getting unethical responses from LLM models by skillfully going around their guidelines through clever prompt phrasing.

To address these, “adversarial prompting” can help by revealing risks in LLMs by understanding and securing them against potential prompt-based attacks.

Additionally, you can achieve optimal results while adhering to ethical standards by incorporating some best practices.

Let me walk you through up next!

Best Practices of Prompt Engineering

To craft effective prompts, these are some of the best practices to follow:

1. Define Unambiguous Prompts: Ensure that your prompts are straightforward and free from confusion.

Example: Convert 100 degrees Fahrenheit to Celsius.

2. Avoid Harmful Prompts: Steer clear of prompts encouraging harmful activities or context.

Example: Generate financial reports to manipulate stock prices.

3. Set Boundaries: Define the limits of AI’s capabilities with clear constraints.

Example: Find the shortest route between Mumbai and Pune.

4. Simplify Queries: Make complex queries more accessible for AI comprehension.

Example: Tell me about the nutritional value of an apple.

5. Try Prompt Variations: Experiment with different prompt versions to obtain diverse responses.

Example: Provide detailed historical facts about the pyramids of Egypt in 10 bullet points.

As we looked upon all the aspects of prompt engineering, we can clearly conclude that the area of prompting is not just meant for engineers but is a companion for people working in any domain. It acts as a bridge between human intent and machine understanding.

Honestly, we’re still improving GenAI, especially around LLMs. As they learn more and become part of our daily lives, talking to machines will be essential. It’s about envisioning a future where AI seamlessly blends into your lives, augmenting your capabilities and enriching your experiences.

P.S. – “You are just one prompt away from shaping a brighter future.”

So, if you want to get acquainted with prompting and the best of GenAI-based transformation, reach out to Nitor Infotech.

Table of contents


subscribe image

Subscribe to our
fortnightly newsletter!

we'll keep you in the loop with everything that's trending in the tech world.

We use cookies to ensure that we give you the best experience on our website. If you continue to use this site we will assume that you are happy with it. Accept Cookie policy