Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Machine Mindset: An MBTI Exploration of Large Language Models

Jiaxi Cui, Liuzhenghao Lv, Jing Wen, Rongsheng Wang, Jing Tang, YongHong Tian, Li Yuan

TL;DR

Problem: personalized LLMs often struggle with stable, interpretable personality expression. Approach: Machine Mindset combines a two-phase supervised fine-tuning on behavior and self-awareness MBTI-aligned datasets with Direct Preference Optimization (DPO), using LoRA adapters to enable modular, switchable personalities in English and Chinese. Contributions: a dataset construction framework for MBTI-aligned behavior and self-awareness, a practical two-stage training regimen, and open-source resources for replication and extension. Impact: provides a data-driven path to stable, domain-specific AI personalization and paves the way for more human-like, context-aware interactions, while acknowledging MBTI limitations and suggesting future multimodal extensions.

Abstract

We present a novel approach for integrating Myers-Briggs Type Indicator (MBTI) personality traits into large language models (LLMs), addressing the challenges of personality consistency in personalized AI. Our method, "Machine Mindset," involves a two-phase fine-tuning and Direct Preference Optimization (DPO) to embed MBTI traits into LLMs. This approach ensures that models internalize these traits, offering a stable and consistent personality profile. We demonstrate the effectiveness of our models across various domains, showing alignment between model performance and their respective MBTI traits. The paper highlights significant contributions in the development of personality datasets and a new training methodology for personality integration in LLMs, enhancing the potential for personalized AI applications. We also open-sourced our model and part of the data at \url{https://github.com/PKU-YuanGroup/Machine-Mindset}.

Machine Mindset: An MBTI Exploration of Large Language Models

TL;DR

Problem: personalized LLMs often struggle with stable, interpretable personality expression. Approach: Machine Mindset combines a two-phase supervised fine-tuning on behavior and self-awareness MBTI-aligned datasets with Direct Preference Optimization (DPO), using LoRA adapters to enable modular, switchable personalities in English and Chinese. Contributions: a dataset construction framework for MBTI-aligned behavior and self-awareness, a practical two-stage training regimen, and open-source resources for replication and extension. Impact: provides a data-driven path to stable, domain-specific AI personalization and paves the way for more human-like, context-aware interactions, while acknowledging MBTI limitations and suggesting future multimodal extensions.

Abstract

We present a novel approach for integrating Myers-Briggs Type Indicator (MBTI) personality traits into large language models (LLMs), addressing the challenges of personality consistency in personalized AI. Our method, "Machine Mindset," involves a two-phase fine-tuning and Direct Preference Optimization (DPO) to embed MBTI traits into LLMs. This approach ensures that models internalize these traits, offering a stable and consistent personality profile. We demonstrate the effectiveness of our models across various domains, showing alignment between model performance and their respective MBTI traits. The paper highlights significant contributions in the development of personality datasets and a new training methodology for personality integration in LLMs, enhancing the potential for personalized AI applications. We also open-sourced our model and part of the data at \url{https://github.com/PKU-YuanGroup/Machine-Mindset}.
Paper Structure (17 sections, 41 figures)

This paper contains 17 sections, 41 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. MBTI: Myers-Briggs Type Indicator
  4. General LLMs
  5. Method
  6. Dataset Construction
  7. Fine-tuning LLMs towards a certain personality
  8. Evaluation Method
  9. Experiments and Results
  10. Training Process
  11. Evaluation Results
  12. Ablation
  13. Effects on Ablities
  14. Conclusion
  15. Details on Experiments
  16. ...and 2 more sections

Figures (41)

  • Figure 1: Machine Mindset Framework
  • Figure 2: The flowchart for constructing behavior datasets and Self-awareness datasets.
  • Figure 3: The ratio of the four MBTI dimensions in behavior datasets
  • Figure 4: MBTI Personality Type Results for INTJ.
  • Figure 5: Second Results for INTJ.
  • ...and 36 more figures