Driving Generative Agents With Their Personality

TL;DR

This paper demonstrates that large language models can effectively incorporate and represent personality profiles, enhancing the realism of game characters by using psychometric data within affective computing systems.

Contribution

It introduces a method for LLMs to utilize psychometric personality data, improving NPC behavior realism in video games, validated through adaptation of the IPIP questionnaire.

Findings

GPT-4 can accurately interpret personality profiles

LLMs enhance NPC human-like behavior

Personality-driven content generation is effective

Abstract

This research explores the potential of Large Language Models (LLMs) to utilize psychometric values, specifically personality information, within the context of video game character development. Affective Computing (AC) systems quantify a Non-Player character's (NPC) psyche, and an LLM can take advantage of the system's information by using the values for prompt generation. The research shows an LLM can consistently represent a given personality profile, thereby enhancing the human-like characteristics of game characters. Repurposing a human examination, the International Personality Item Pool (IPIP) questionnaire, to evaluate an LLM shows that the model can accurately generate content concerning the personality provided. Results show that the improvement of LLM, such as the latest GPT-4 model, can consistently utilize and interpret a personality to represent behavior.