Problem
Research questions and friction points this paper is trying to address.
Metacognition
Large Language Models
Artificial Intelligence
Cognitive Abilities
AI Reliability
Metacognition in LLMs: Foundations, Progress, and Opportunities
🤖 AI Summary
Whether large language models possess metacognitive capabilities—and how to effectively elicit and leverage such abilities to enhance the intelligence, reliability, and transparency of AI systems—remains an open question. This work presents the first comprehensive survey and systematic taxonomy of metacognition in large language models, synthesizing evaluation benchmarks, elicitation and enhancement techniques, and representative application scenarios. Through extensive literature analysis, technical categorization, and illustrative case studies, the paper constructs a structured knowledge framework and releases an open-source bibliography to provide a foundational theoretical basis, methodological roadmap, and clear directions for future research in this emerging field.
📝 Abstract
Metacognition is a foundational component of intelligence critical to effective learning, problem solving, decision-making, communication, and more. In recent years, it has become increasingly recognized as a cornerstone of capable, transparent AI systems. Yet while LLMs have made significant progress across diverse real-world tasks, it is not yet clear when, how, or to what extent they can exhibit or be endowed with effective metacognitive abilities, nor how such abilities can be adapted to advance the fundamental capabilities, reliability, and intelligence of AI systems. This paper bridges this gap by presenting the first comprehensive overview of the current state of knowledge on metacognition for LLMs. We analyze and taxonomize the landscape of this emerging field and summarize recent technical advancements, including methods and benchmarks to measure and evaluate LLMs' metacognitive abilities, techniques to elicit, improve, and apply metacognition in LLMs, and findings and implications of ongoing research. We also discuss applications, open questions and challenges, and promising directions for future work. Our aim is to provide a detailed and up-to-date review of this topic and stimulate meaningful research and discussion. An organized list of papers can be found at https://github.com/yale-nlp/LLM-Metacognition.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Metacognition
Large Language Models
AI Reliability
Cognitive Evaluation
Self-monitoring
🔎 Similar Papers
No similar papers found.
