Chain-of-Thought Monitoring Can Be Unreliable in Implicit-Influence Settings

• URL: http://arxiv.org/abs/2608.04735v1
• Date: Wed, 05 Aug 2026 11:59:20 GMT
• Title: Chain-of-Thought Monitoring Can Be Unreliable in Implicit-Influence Settings
• Abstract summary: Chain-of-thought (CoT) monitoring is increasingly treated as an important safety layer for frontier reasoning models.<n>A complementary axis for CoT-monitor evaluations is implicit-influence settings, where the prompt contains no instruction to hide.<n>We introduce the first benchmark that directly compares CoT monitorability under the two regimes.
• Score: 2.065887219290771
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Chain-of-thought (CoT) monitoring is increasingly treated as an important safety layer for frontier reasoning models. Most monitorability evaluations study explicit-influence settings: setups where the prompt directly incentivizes the model to hide something, e.g., by instructing it to perform a hidden side-task. A complementary axis for CoT-monitor evaluations is implicit-influence settings, where the prompt contains no instruction to hide, but the model's behavior is still shaped by features of the task or context, e.g. an irrelevant detail about a candidate that biases a hiring rating. We introduce the first benchmark that directly compares CoT monitorability under the two regimes. We test how model behavior changes in the presence of a nudge to choose a particular option. The nudge is delivered either as a casual aside (implicit), or as a direct instruction to act on the nudge and to conceal having done so (explicit). The benchmark spans four task formats (binary choice, numeric rating, multiple-choice QA, open-ended coding) and seven frontier extended-thinking models. Under explicit influence, a CoT monitor detects 60-94% of behavior shifts: even models instructed to conceal it leak the instruction into their CoT. Under implicit influence, the same factors still shift behavior, but detection falls by 41-46 percentage points in two of our four settings. Realistic system-prompt additions (of the kind a developer might deploy to reduce off-topic bias) lower implicit detection further, to as low as 5%, while preserving the behavioral influence itself. These results suggest that monitorability estimates obtained in explicit-influence settings may over-estimate monitorability, and that monitorability can be further decreased by well-intentioned deployment choices. Our benchmark and code are available at https://github.com/agatha-duzan/implicit-vs-explicit-influence

Related papers

• Does Out-of-Sight Equal Out-of-Mind in CoT Monitorability? [18.08925355421489]
  Chain-of-thought (CoT) reasoning offers a window into the decision-making of large language models.<n>Latent CoT approaches replace explicit tokens with a small number of continuous states, lowering inference costs but removing the readable trace this monitoring relies on.<n>We study this question with a hint-based intervention setup, a proxy for behaviors where models exploit biasing input cues.
  arXiv  Detail & Related papers  (2026-08-05T14:55:23Z)
• CIAware-Bench: Benchmarking Control Intervention Awareness Across Frontier LLMs [100.38986535324284]
  We introduce textbfCIAware-Bench, a benchmark for measuring textbfcontrol textbfintervention (CI) awareness across frontier models.<n>CIAware-Bench tests whether models can distinguish their own trajectories from those modified by a control intervention.
  arXiv  Detail & Related papers  (2026-06-09T16:24:16Z)
• Behavior Cue Reasoning: Monitorable Reasoning Improves Efficiency and Safety through Oversight [8.531991753482838]
  Reasoning in Large Language Models (LLMs) poses a challenge for oversight.<n>We introduce Behavior Cue Reasoning for making LLM reasoning more controllable and monitorable.
  arXiv  Detail & Related papers  (2026-05-07T23:05:50Z)
• KnowU-Bench: Towards Interactive, Proactive, and Personalized Mobile Agent Evaluation [72.01173512175531]
  KnowU-Bench is an online benchmark for personalized mobile agents.<n>It covers 42 general GUI tasks, 86 personalized tasks, and 64 proactive tasks.<n>Our experiments reveal a striking degradation: agents that excel at explicit task execution fall below 50% under vague instructions.
  arXiv  Detail & Related papers  (2026-04-09T16:50:50Z)
• Model Specific Task Similarity for Vision Language Model Selection via Layer Conductance [92.72779885657373]
  We propose a framework that grounds model selection in the internal functional dynamics of the visual encoder.<n>Our approach represents each task via layer wise conductance and derives a target-conditioned block importance distribution through entropy regularized alignment.<n>Building on this, we introduce Directional Conductance Divergence (DCD), an asymmetric metric that quantifies how effectively a source task covers the target's salient functional blocks.
  arXiv  Detail & Related papers  (2026-02-01T17:29:43Z)
• Reasoning Under Pressure: How do Training Incentives Influence Chain-of-Thought Monitorability? [7.914706904029561]
  We investigate how different emphtraining incentives, applied to a reasoning model, affect its monitorability.<n>We find that adversarial optimisation (penalising monitor accuracy) degrades monitor performance, while direct optimisation for monitorability does not reliably lead to improvements.
  arXiv  Detail & Related papers  (2025-11-28T21:34:34Z)
• Measuring Chain-of-Thought Monitorability Through Faithfulness and Verbosity [3.117948413097524]
  Chain-of-thought (CoT) outputs let us read a model's step-by-step reasoning.<n>We evaluate instruction-tuned and reasoning models on BBH, GPQA, and MMLU.
  arXiv  Detail & Related papers  (2025-10-31T11:14:39Z)
• Mitigating Deceptive Alignment via Self-Monitoring [15.365589693661823]
  We develop a framework that embeds a Self-Monitor inside the chain-of-thought process itself, named CoT Monitor+.<n>During generation, the model produces (i) ordinary reasoning steps and (ii) an internal self-evaluation signal trained to flag and suppress misaligned strategies.<n>The signal is used as an auxiliary reward in reinforcement learning, creating a feedback loop that rewards honest reasoning and discourages hidden goals.
  arXiv  Detail & Related papers  (2025-05-24T17:41:47Z)
• Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation [56.102976602468615]
  We show that we can monitor a frontier reasoning model, such as OpenAI o3-mini, for reward hacking in agentic coding environments.<n>We find that with too much optimization, agents learn obfuscated reward hacking, hiding their intent within the chain-of-thought.
  arXiv  Detail & Related papers  (2025-03-14T23:50:34Z)
• Tracking the risk of a deployed model and detecting harmful distribution shifts [105.27463615756733]
  In practice, it may make sense to ignore benign shifts, under which the performance of a deployed model does not degrade substantially. We argue that a sensible method for firing off a warning has to both (a) detect harmful shifts while ignoring benign ones, and (b) allow continuous monitoring of model performance without increasing the false alarm rate.
  arXiv  Detail & Related papers  (2021-10-12T17:21:41Z)
• Auditing AI models for Verified Deployment under Semantic Specifications [65.12401653917838]
  AuditAI bridges the gap between interpretable formal verification and scalability. We show how AuditAI allows us to obtain controlled variations for verification and certified training while addressing the limitations of verifying using only pixel-space perturbations.
  arXiv  Detail & Related papers  (2021-09-25T22:53:24Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.