Simulating phenomenal consciousness using generative agents based on large language models

Han‐Zhong Zhang

doi:10.1016/j.asoc.2025.113922

Simulating phenomenal consciousness using generative agents based on large language models

2025-09-22

0

PRE

AI

DOI：10.1016/j.asoc.2025.113922

Original

Original request for help

Save

Abstract

En 中文

• To enhance the foundational structure of generative agents, making them more intelligent and interpretable, we propose the Internal Time-Consciousness Machine (ITCM) as a reconstruction of the underlying architecture of generative agents. This computational consciousness structure can assist generative agents in handling complex tasks with greater flexibility and intelligence while enhancing their interpretability, making their behaviors easier to understand and predict. • Technically, based on the proposed ITCM structure, we introduce an ITCM-based agent (ITCMA). • Evaluations in the Alfworld environment show that trained ITCMA outperforms the state-of-the-art (SOTA) by 9 % on the seen set. Even untrained ITCMA achieves a 96 % task completion rate on the seen set, 5 % higher than SOTA, indicating its superiority over traditional intelligent agents in utility and generalization. • In real-world tasks with quadruped robots, the untrained ITCMA achieves an 85 % task completion rate, which is close to its performance in the unseen set, demonstrating its comparable utility in real-world settings.

Simulating phenomenal consciousness using generative agents based on large language models

Abstract

Journal

Organization

Cited Papers