PandaVT/Machine_Mindset_MBTI_dataset

Dataset Introduction

This dataset is for supervised fine‑tuning (SFT) and direct preference optimization (DPO). It is available in English (prefix en) and Chinese (prefix zh).

MBTI Dimensions

The dataset is based on the four MBTI dimensions, each containing two opposing attributes:

• Energy: Extraversion (E) – Introversion (I)
• Information: Sensing (S) – Intuition (N)
• Decision: Thinking (T) – Feeling (F)
• Execution: Judging (J) – Perceiving (P)

Data Format

The dataset follows the Alpaca format, containing instruction, input, and output.

Usage

Supervised Fine‑Tuning (SFT)

For example, to endow an LLM with ISFJ traits, select the following four corresponding files for SFT:

• en_energe_introversion.json
• en_information_sensing.json
• en_decision_feeling.json
• en_execution_judging.json

Direct Preference Optimization (DPO)

For instance, to bias an LLM toward Feeling (F) rather than Thinking (T) via DPO, select the following two corresponding files:

• en_decision_feeling.json
• en_decision_thinking.json

Then compile them into the correct DPO format.