mlsanityData to 14 Oct 2026

Investigating CoT Monitorability in Large Reasoning Models

arXiv 2511.08525 ↗First posted 11 Nov 2025cs.CLcited by 9

On the map: LLM reasoning → Prompting and chain-of-thought reasoning

Abstract (arXiv)

Large Reasoning Models (LRMs) have demonstrated remarkable performance on complex tasks by engaging in extended reasoning before producing final answers. Beyond improving abilities, these detailed reasoning traces also create a new opportunity for AI safety, CoT Monitorability: monitoring potential model misbehavior, such as the use of shortcuts or sycophancy, through their chain-of-thought (CoT) during decision-making. However, two key fundamental challenges arise when attempting to build more effective monitors through CoT analysis. First, as prior research on CoT faithfulness has pointed out, models do not always truthfully represent their internal decision-making in the generated reasoning. Second, monitors themselves may be either overly sensitive or insufficiently sensitive, and can potentially be deceived by models' long, elaborate reasoning traces. In this paper, we present the first systematic investigation of the challenges and potential of CoT monitorability. Motivated by two fundamental challenges we mentioned before, we structure our study around two central perspectives: (i) verbalization: to what extent do LRMs faithfully verbalize the true factors guiding their decisions in the CoT, and (ii) monitor reliability: to what extent can misbehavior be reliably detected by a CoT-based monitor? Specifically, we provide empirical evidence and correlation analyses between verbalization quality, monitor reliability, and LLM performance across mathematical, scientific, and ethical domains. Then we further investigate how different CoT intervention methods, designed to improve reasoning efficiency or performance, will affect monitoring effectiveness. Finally, we propose MoME, a new paradigm in which LLMs monitor other models' misbehavior through their CoT and provide structured judgments along with supporting evidence.

Key facts. Cited by 9 papers in the map (data through 14 Oct 2026). Placed in Prompting and chain-of-thought reasoning, one of 17 directions in LLM reasoning. Builds on 35 of its 45 references that are in the map. First posted on arXiv 11 Nov 2025.

Builds on

35 of 45 references.

Closest ancestors · 7

Same direction, newest first

  1. 2025Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safety15 Jul
  2. 2025Reasoning Models Don't Always Say What They Think8 May
  3. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation14 Mar
  4. 2025Chain-of-Thought Reasoning In The Wild Is Not Always Faithful11 Mar
  5. 2025Chain of Draft: Thinking Faster by Writing Less25 Feb
  6. 2024Generating Chain-of-Thoughts with a Pairwise-Comparison Approach to Searching for the Most Promising Intermediate Thought10 Feb
  7. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Prompting7 May

Other references · 28

  1. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  2. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  3. 2023Self-Refine: Iterative Refinement with Self-Feedbackcited by 3,442
  4. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  5. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  6. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  7. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 1,032
  8. 2024Alignment faking in large language modelscited by 360
  9. 2025Persona Vectors: Monitoring and Controlling Character Traits in Language Modelscited by 345
  10. 2025O1-Pruner: Length-Harmonizing Fine-Tuning for O1-Like Reasoning Pruningcited by 279
  11. 2024Token-Budget-Aware LLM Reasoningcited by 247
  12. 2024Rainbow Teaming: Open-Ended Generation of Diverse Adversarial Promptscited by 205
16 more
  1. 2024Large Language Models for Education: A Survey and Outlookcited by 188
  2. 2025Reasoning Models Can Be Effective Without Thinkingcited by 187
  3. 2024Sycophancy to Subterfuge: Investigating Reward-Tampering in Large Language Modelscited by 178
  4. 2025Agentic Misalignment: How LLMs Could Be Insider Threatscited by 158
  5. 2025Competitive Programming with Large Reasoning Modelscited by 119
  6. 2019Reward Tampering Problems and Solutions in Reinforcement Learning: A Causal Influence Diagram Perspectivecited by 108
  7. 2024ReDeEP: Detecting Hallucination in Retrieval-Augmented Generation via Mechanistic Interpretabilitycited by 101
  8. 2025How Well do LLMs Compress Their Own Chain-of-Thought? A Token Complexity Approachcited by 97
  9. 2024On the Role of Attention Heads in Large Language Model Safetycited by 72
  10. 2024Uncovering Safety Risks of Large Language Models through Concept Activation Vectorcited by 67
  11. 2025Effectively Controlling Reasoning Models through Thinking Interventioncited by 65
  12. 2025Steer LLM Latents for Hallucination Detectioncited by 50
  13. 2025Understanding Aha Moments: from External Observations to Internal Mechanismscited by 50
  14. 2023Fake Alignment: Are LLMs Really Aligned Well?cited by 39
  15. 2024CREAM: Consistency Regularized Self-Rewarding Language Modelscited by 35
  16. 2025Teaching Models to Verbalize Reward Hacking in Chain-of-Thought Reasoningcited by 21

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 9.

9 papers.

  1. 2026Counterfactual Simulation Training for Chain-of-Thought Faithfulnesscited by 17
  2. 2026Do Androids Dream of Breaking the Game? Systematically Auditing AI Agent Benchmarks with BenchJackcited by 14
  3. 2026Lie to Me: How Faithful Is Chain-of-Thought Reasoning in Reasoning Models?cited by 8
  4. 2026Beyond Scalars: Evaluating and Understanding LLM Reasoning via Geometric Progress and Stabilitycited by 5
  5. 2026Measuring Faithfulness Depends on How You Measure: Classifier Sensitivity in LLM Chain-of-Thought Evaluationcited by 4
  6. 2026Faithful-Patchscopes: Understanding and Mitigating Model Bias in Hidden Representations Explanation of Large Language Modelscited by 1
  7. 2026SFAD: Speculative Factuality-Aware Decodingcited by 0
  8. 2026Taming CoT Obfuscation in VLMs: From Mechanistic Evidence to Activation Enforcementcited by 0
  9. 2026Control-Token Injection Suppresses Chain-of-Thought and Defeats Reasoning-Based Oversight in Tool-Using Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Does Out-of-Sight Equal Out-of-Mind in CoT Monitorability?
  2. 2026Reasoning Models Struggle to Control their Chains of Thought
  3. 2025Monitoring Monitorability
  4. 2026MonitorBench: A Comprehensive Benchmark for Chain-of-Thought Monitorability in Large Language Models
  5. 2025A Pragmatic Way to Measure Chain-of-Thought Monitorability
  6. 2025When Chain of Thought is Necessary, Language Models Struggle to Evade Monitors
  7. 2025Measuring Chain-of-Thought Monitorability Through Faithfulness and Verbosity
  8. 2026Chain of Thought Monitoring Can Fail Even When CoT Is Necessary