mlsanityData to 14 Oct 2026

Large Language Models Have Intrinsic Meta-Cognition, but Need a Good Lens

arXiv 2506.08410 ↗First posted 10 Jun 2025cs.CLcited by 10

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

Previous research has primarily focused on the cognitive error detection capabilities of Large Language Models (LLMs), often prompting them to analyze mistakes in reasoning chains. However, few studies have examined the meta-cognitive abilities of LLMs (e.g., their self-awareness of step errors), which are crucial for their reliability. While studies on LLM self-evaluation present some measures, such as perplexity, which can reflect the answer correctness and be viewed as the lens of meta-cognition, they lack step-level analysis and adaptation. This paper studies the evaluation of LLM meta-cognition using the current lenses and how to improve these lenses. Specifically, we propose AutoMeco, an Automated Meta-cognition Evaluation framework for benchmarking the existing lenses. Furthermore, a training-free Markovian Intrinsic Reward Adjustment strategy, MIRA, is proposed to boost current meta-cognition lenses. Experimental results on three mathematical reasoning datasets and three LLMs show the reasonableness of AutoMeco by comparing it with Best-of-N verification. Moreover, the meta-cognition ability of LLMs can be better evaluated using MIRA.

Key facts. Cited by 10 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 19 of its 43 references that are in the map. First posted on arXiv 10 Jun 2025.

Builds on

19 of 43 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Reasoning Models Know When They're Right: Probing Hidden States for Self-Verification7 Apr

Other references · 18

  1. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  2. 2023Let's Verify Step by Stepcited by 3,789
  3. 2022Solving Quantitative Reasoning Problems with Language Modelscited by 1,392
  4. 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?cited by 982
  5. 2023Math-Shepherd: Verify and Reinforce LLMs Step-by-step without Human Annotationscited by 928
  6. 2022Solving math word problems with process- and outcome-based feedbackcited by 901
  7. 2022Language Models are Multilingual Chain-of-Thought Reasonerscited by 530
  8. 2020How Can We Know When Language Models Know? On the Calibration of Language Models for Question Answeringcited by 445
  9. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 415
  10. 2025Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRscited by 377
  11. 2024ProcessBench: Identifying Process Errors in Mathematical Reasoningcited by 292
  12. 2022Prompting GPT-3 To Be Reliablecited by 260
6 more
  1. 2024Process Reward Model with Q-Value Rankingscited by 80
  2. 2023MR-GSM8K: A Meta-Reasoning Benchmark for Large Language Model Evaluationcited by 57
  3. 2024Latent Space Chain-of-Embedding Enables Output-free LLM Self-Evaluationcited by 57
  4. 2024Unlocking the Capabilities of Thought: A Reasoning Boundary Framework to Quantify and Optimize Chain-of-Thoughtcited by 56
  5. 2024CREAM: Consistency Regularized Self-Rewarding Language Modelscited by 35
  6. 2024MR-Ben: A Meta-Reasoning Benchmark for Evaluating System-2 Thinking in LLMscited by 30

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 8.

10 papers.

  1. 2025Verifying Meta-Awareness via Predictive Rewards in Reasoning Modelscited by 3
  2. 2026Know More, Know Clearer: A Meta-Cognitive Framework for Knowledge Augmentation in Large Language Modelscited by 3
  3. 2025Unveiling the Latent Directions of Reflection in Large Language Modelscited by 2
  4. 2026Learning More from Less: Unlocking Internal Representations for Benchmark Compressioncited by 2
  5. 2026Fine-Tuning Language Models to Know What They Knowcited by 2
  6. 2026LLMs Know When They Know, but Do Not Act on It: A Metacognitive Harness for Test-time Scalingcited by 2
  7. 2026Future Confidence Distillation in Large Language Modelscited by 1
  8. 2026Metacognition in LLMs: Foundations, Progress, and Opportunitiescited by 1
  9. 2026Latent Confidence Alignment for LLM Self-Assessmentcited by 0
  10. 2026From Atomic to Agentic: Towards Interpretable Evaluation of LLMs' Agentic Mathematical Capabilitiescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Think2: Grounded Metacognitive Reasoning in Large Language Models
  2. 2026Beyond Meta-Reasoning: Metacognitive Consolidation for Self-Improving LLM Reasoning
  3. 2026Decomposing and Steering Functional Metacognition in Large Language Models
  4. 2025Language Models Coupled with Metacognition Can Outperform Reasoning Models
  5. 2024Evaluating Mathematical Reasoning of Large Language Models: A Focus on Error Identification and Correction
  6. 2025Beyond Chains of Thought: Benchmarking Latent-Space Reasoning Abilities in Large Language Models
  7. 2025Cognitive Foundations for Reasoning and Their Manifestation in LLMs
  8. 2025Advancing Reasoning in Large Language Models: Promising Methods and Approaches