Can someone suggest the correct Prompt?

Hi All,

I am trying to write prompt which enables LLM to provide answers on complicated blood relation puzzles. I tried several prompts but could not write a prompt which gives me correct response. Please any help is appreciated.

I am using prompt like:
Act as a complex problem solver and analyze the text given below.
Consider all the relation and job details provided in the text.
Create family tree with job details
Check for couples.
Check for siblings.
Cross check the tree again and create understanding between the relationship among all the people in the text against the information provided.
Question: How Neerja is related to Vinita?
The text I am using:
When Vina Patni, tops in her Medical course, she becomes the 7th member of her family to join the Medical profession. All other family members namely Tanmay, Neerja, Monnish, Sanchit, Ruchika and Vinita are working at Sawai Man Singh Hospital as doctors and Vina who is the youngest of the third generation, is all set to join them. In this family of three generations, there are two pairs of couples and two pairs of brothers and sisters. There are three Oncologists, two Dentists, one Physician and one Neurologist in the family. No female in the family is an Oncologist. Vinita is not a Neurologist. No married person in the family is a Dentist. Tanmay is the son of Neerja while Sanchit is the son of Monnish and Monnish’s wife is Vinita. Monnish is an Oncologist. Monnish and Vinita are the eldest couple in the family. Vina is the daughter of Sanchit while Ruchika is the sister-in-law of Neerja.

Correct Answer: Daughter-In-Law.
Answer I am getting: Sister-In-Law

Hi @Garvita

Welcome to the community.

What course are you referring to? Notice that you posted on the general category.

It is very important to post on the proper space in order to get a proper support. This is the only way mentors could be aware of learners queries and try to help.

Don’t forget to check our Guidelines. It is a good started point.

Best regards


I am referring to course by openAI and on prompt engineering " [ChatGPT Prompt Engineering for Developers - DeepLearning.AI]ChatGPT Prompt Engineering for Developers - DeepLearning.AI".

Though, there was no such exercise mentioned in the course and I was doing some hand on using prompts for various types of problems. And hence, posted the query in general category.
I have updated the category.

Sure, I will go through the guidelines.

Hi @Garvita

Thanks for the update.

This is a expect behavior regards LLMs. Keep in mind that the ChatGPT it is a language model. It was build it with a large language dataset. The algorithm used to training this models is based on a probabilistic approach. It calculates the probabilistic to a certain word comes after another. It is like your auto text corrector, but with steroids. That means that the LLMs doesn’t provide a “logic thinking” feature.

To achieve a better results, your prompt must “explain” what a family tree is and how this tree is structured. You have to “explain” all the relationships between every relative and so on.

Even improving your prompt following this tips doesn’t guarantee that the model will give you the right answer.

I hope this help.

Best regards

Hi there i wrote down below prompts which have same meaning as usuall and got 2 different answers and i am wonder what is the difference. and why i got diffrent answer for chatgpt?

lamp_review = “”"
Needed a nice lamp for my bedroom, and this one had
additional storage and not too high of a price point.
Got it fast. The string to our lamp broke during the
transit and the company happily sent over a new one.
Came within a few days as well. It was easy to put
together. I had a missing part, so I contacted their
support and they very quickly got me the missing piece!
Lumina seems to me to be a great company that cares
about their customers and products!!
prompt 1= f"""is the writer of following review express negative emotions?
Identify a negative emotions that writer of the following review was expressing,
which is delimited with triple backticks.

Review text: {lamp_review}

prompt 2= f"""Identify a negative emotions that writer of the following review was expressing,
which is delimited with triple backticks.

Review text: {lamp_review}