CogGPT: Unleashing the Power of Cognitive Dynamics on Large Language Models
This work addresses the gap in dynamic cognitive simulation for LLMs, which is incremental as it builds on existing static approaches to improve modeling for AI and cognitive science applications.
The paper tackles the problem of static cognitive modeling in large language models (LLMs) by introducing the concept of cognitive dynamics, proposing a benchmark (CogBench) with evaluation metrics, and developing CogGPT with an iterative mechanism to enhance lifelong cognitive dynamics, showing superiority over existing methods in facilitating role-specific dynamics under continuous information flows.
Cognitive dynamics are pivotal to advance human understanding of the world. Recent advancements in large language models (LLMs) reveal their potential for cognitive simulation. However, these LLM-based cognitive studies primarily focus on static modeling, overlooking the dynamic nature of cognition. To bridge this gap, we propose the concept of the cognitive dynamics of LLMs and present a corresponding task with the inspiration of longitudinal studies. Towards the task, we develop CogBench, a novel benchmark to assess the cognitive dynamics of LLMs and validate it through participant surveys. We also design two evaluation metrics for CogBench, including Authenticity and Rationality. Recognizing the inherent static nature of LLMs, we introduce CogGPT for the task, which features an innovative iterative cognitive mechanism aimed at enhancing lifelong cognitive dynamics. Empirical results demonstrate the superiority of CogGPT over existing methods, particularly in its ability to facilitate role-specific cognitive dynamics under continuous information flows.