Persona-Code

Introduction

Inspired by research that links task-personality alignment with improved development outcomes, we conduct an empirical study on personality-guided code generation using large language models (LLMs). This repository is the code and data of this study.

Directory Structure

• MBPPGen - Contains core code and resources related to personality-guided code generation for MBPP Sanitized
• HumanEvalGen- Contains core code and resources related to personality-guided code generation for HumanEval +
• MBPP_Plus - Contains core code and resources related to personality-guided code generation for MBPP +
• APPSGen - Contains core code and resources related to personality-guided code generation for APPS

Installation

Before using this project, ensure you have installed all necessary dependencies. You can follow these steps to install them:

1. Clone this repository:

   https://github.com/ab3xyz/Persona-Code.gitgit clone <repository-url>
   cd Persona-Code

2. Install dependencies:

   pip install -r requirements.txt

Usage

Running Code Generation

First add your own API_KEY to the specific model, and modify the experiment_<model>.py in directory of the datasets to run the test you want.

Data Generated

See the data_<model> part for data generated by this study.