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

TL;DR

This paper introduces a metacognitive framework inspired by psychological theory to improve large language models' ability to monitor, diagnose, and correct their errors, significantly enhancing their self-correction and trustworthiness.

Contribution

It operationalizes a psychological regulatory cycle as a prompting architecture and integrates it into a dual-process MetaController, improving error diagnosis and self-correction in LLMs.

Findings

Threefold increase in successful self-correction.

84% preference for trustworthiness in human evaluations.

Significant improvement across multiple reasoning benchmarks.

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.