Modeling Layered Consciousness with Multi-Agent Large Language Models

TL;DR

This paper introduces a multi-agent framework grounded in psychoanalytic theory to model artificial consciousness in large language models, demonstrating improved emotional depth and personalization.

Contribution

It presents a novel psychodynamic multi-agent model for LLMs that simulates consciousness layers and enhances personalization through fine-tuning and agent interaction.

Findings

71.2% preference for fine-tuned model in evaluations

Improved emotional depth in generated outputs

Reduced output variance

Abstract

We propose a multi-agent framework for modeling artificial consciousness in large language models (LLMs), grounded in psychoanalytic theory. Our \textbf{Psychodynamic Model} simulates self-awareness, preconsciousness, and unconsciousness through agent interaction, guided by a Personalization Module combining fixed traits and dynamic needs. Using parameter-efficient fine-tuning on emotionally rich dialogues, the system was evaluated across eight personalized conditions. An LLM as a judge approach showed a 71.2\% preference for the fine-tuned model, with improved emotional depth and reduced output variance, demonstrating its potential for adaptive, personalized cognition.