Introduction
This study aims to investigate whether LLMs can answer some questions about human MBTI. The study will answer questions based on the responses from LLMs. There are three research questions in this research:
Do LLMs understand MBTI?
Exploring the comprehension capabilities of Large Language Models regarding Myers-Briggs Type Indicator concepts.
Could LLMs think about the personalities from a human perspective?
Investigating whether LLMs can simulate human-like personality reasoning.
How accurate of the LLMs' responses in MBTI?
Measuring the precision and reliability of LLM-generated MBTI-related responses.
Research Team
Methodology
Question Design
Devised 17 MBTI-related questions for evaluation
LLM Testing
Presented questions to 10 different Large Language Models
Data Collection
Compiled responses into Excel files for ENFP and INTJ types
Analysis
Used Python for data visualization and analysis
This study first devised 17 MBTI-related questions, which were then presented by ten large language models (LLMs) to pretend they were the people with MBTI: ENFP/INTJ (each model consisted of 17 questions). The responses were then compiled into two Excel files. Python was used to create bar charts to visualize the data. Finally, the data was analyzed into two parts:
Part A
What are the same/different answers on INTJ and ENFP personalities with different LLMs?
Part B
What are the LLMs' views on INTJ and ENFP personalities?
Download Research Data
Click on any Excel file below to download the complete research data
ENFP short Responses
Short LLM responses for ENFP personality type
ENFP Long Responses
Long LLM responses for ENFP personality type
INTJ Short responses
Short LLM responses for INTJ personality type
INTJ Long response
Long LLM responses for INTJ personality type
All Excel files contain raw research data. Files include response timestamps, model information, and complete question-answer pairs for reproducibility.
Acknowledgements
| Members | Contributions |
|---|---|
|
Lin Ka Man
55481716 |
Correcting questionnaire questions, Integrating LLMs' responses (INTJ), Analyze Part B, Figures 1-17, The content of introduction and methodology |
|
Cheung Mung Fei
58073283 |
Creating questionnaire questions, Integrating LLMs' responses (ENFP), Analyze Part A, Figures 18-24, Making the website, simple word doc about web design, zip file |
Declaration of GenAI Use
Data Visualization
Used LLMs for creating visual representations of research data and analysis results.
Website Development
Utilized AI tools for website structure, design implementation, and code optimization.
This project transparently incorporates AI tools to enhance research presentation while maintaining academic integrity.