Large Language Models as General Purpose Intelligence Systems for Reasoning, Planning and Decision Making
Fengyuan Zhang, Bi Wu
Abstract
The emergence of large language models (LLMs) has fundamentally transformed artificial intelligence (AI) research and applications, positioning these systems as potential candidates for general purpose intelligence. Large language models are deep neural networks trained on massive text corpora that demonstrate remarkable capabilities across diverse cognitive tasks without task-specific fine-tuning. This review examines how LLMs function as general intelligence systems, with particular emphasis on three core cognitive domains: reasoning, planning, and decision making. We analyze the architectural foundations that enable LLMs to perform complex reasoning tasks, including chain-of-thought prompting (CoT), in-context learning (ICL), and emergent abilities that arise from scale. The planning capabilities of LLMs are evaluated through their performance on multi-step problem decomposition, goal-oriented task completion, and strategic action sequencing. Furthermore, we investigate decision-making frameworks where LLMs serve as autonomous agents, policy advisors, and collaborative systems that integrate human expertise with machine intelligence. The review synthesizes recent advances in prompt engineering, retrieval-augmented generation (RAG), and multimodal integration that enhance LLM capabilities for general intelligence tasks. We examine real-world applications spanning healthcare diagnosis, financial analysis, scientific discovery, and autonomous systems management. Critical challenges including hallucination, reasoning consistency, computational efficiency, and ethical considerations are thoroughly discussed. This comprehensive analysis demonstrates that while LLMs exhibit significant progress toward general purpose intelligence, fundamental limitations in causal understanding, long-term planning coherence, and adaptive learning remain open research challenges that require continued innovation in architecture design, training methodologies, and evaluation frameworks.