Think$^{2}$: Grounded Metacognitive Reasoning in Large Language Models

2026-02-21

AI Summary

This work addresses the lack of effective metacognitive mechanisms in large language models (LLMs) for self-monitoring and error correction. It introduces, for the first time, Ann Brown’s cognitive regulation loop theory into LLM reasoning through a novel adaptive framework that integrates structured prompting with a lightweight dual-process MetaController, enabling interpretable and diagnosable self-correction. Experiments on Llama-3 and Qwen-3 (8B) demonstrate that the proposed approach triples self-correction success rates across multiple reasoning and diagnostic benchmarks compared to standard methods. Furthermore, in human evaluations, 84% of participants rated the model’s outputs as significantly more trustworthy and reflective of metacognitive awareness than those generated by conventional reasoning or chain-of-thought approaches.

Technology Category

Machine Learning: Large Multimodal Models (LMMs)Search and Optimization: Metareasoning and MetaheuristicsCognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendationEconomics, Online Markets and Human Computation: LLM based quality controls for crowd workSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Large Language Models (LLMs) demonstrate strong reasoning performance, yet their ability to reliably monitor, diagnose, and correct their own errors remains limited. We introduce a psychologically grounded metacognitive framework that operationalizes Ann Brown's regulatory cycle (Planning, Monitoring, and Evaluation) as a structured prompting architecture, and study its integration within a lightweight dual-process MetaController for adaptive effort allocation. Across diverse reasoning and diagnostic benchmarks (GSM8K, CRUXEval, MBPP, AIME, CorrectBench, and TruthfulQA) using Llama-3 and Qwen-3 (8B), explicit regulatory structuring substantially improves error diagnosis and yields a threefold increase in successful self-correction. Blinded human evaluations over 580 query pairs show an 84% aggregate preference for trustworthiness and metacognitive self-awareness over standard and Chain-of-Thought baselines. Grounding LLM reasoning in established cognitive theory offers a principled path toward more transparent and diagnostically robust AI systems.

Problem

• metacognitive reasoning
• error diagnosis
• self-correction
• large language models
• cognitive regulation

Innovation

• metacognitive reasoning
• regulatory cycle
• self-correction
• structured prompting
• dual-process MetaController