mlsanityData to 14 Oct 2026

Difficulties with Evaluating a Deception Detector for AIs

arXiv 2511.22662 ↗First posted 27 Nov 2025cs.LGcited by 13

Not on the map. Why?

Abstract (arXiv)

Building reliable deception detectors for AI systems -- methods that could predict when an AI system is being strategically deceptive without necessarily requiring behavioural evidence -- would be valuable in mitigating risks from advanced AI systems. But evaluating the reliability and efficacy of a proposed deception detector requires examples that we can confidently label as either deceptive or honest. We argue that we currently lack the necessary examples and further identify several concrete obstacles in collecting them. We provide evidence from conceptual arguments, analysis of existing empirical works, and analysis of novel illustrative case studies. We also discuss the potential of several proposed empirical workarounds to these problems and argue that while they seem valuable, they also seem insufficient alone. Progress on deception detection likely requires further consideration of these problems.

Key facts. Cited by 13 papers in the map (data through 14 Oct 2026). Builds on 51 of its 90 references that are in the map. First posted on arXiv 27 Nov 2025.

Builds on

51 of 90 references.

References · 51

  1. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  2. 2022Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedbackcited by 3,720
  3. 2016Concrete Problems in AI Safetycited by 1,642
  4. 2016Understanding intermediate layers using linear classifier probescited by 1,353
  5. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,336
  6. 2023Steering Language Models With Activation Engineeringcited by 922
  7. 2024Training Large Language Models to Reason in a Continuous Latent Spacecited by 797
  8. 2022Discovering Latent Knowledge in Language Models Without Supervisioncited by 767
  9. 2023The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetscited by 712
  10. 2023The Internal State of an LLM Knows When It's Lyingcited by 659
  11. 2024The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearningcited by 509
  12. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
38 more
  1. 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 422
  2. 2018AI safety via debatecited by 400
  3. 2025Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscationcited by 384
  4. 2024Alignment faking in large language modelscited by 360
  5. 2024Frontier Models are Capable of In-context Schemingcited by 287
  6. 2019Risks from Learned Optimization in Advanced Machine Learning Systemscited by 258
  7. 2022The Alignment Problem from a Deep Learning Perspectivecited by 258
  8. 2024OR-Bench: An Over-Refusal Benchmark for Large Language Modelscited by 244
  9. 2022Talking About Large Language Modelscited by 175
  10. 2025Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMscited by 172
  11. 2025Chain-of-Thought Reasoning In The Wild Is Not Always Faithfulcited by 170
  12. 2023Large Language Models can Strategically Deceive their Users when Put Under Pressurecited by 146
  13. 2024AI Sandbagging: Language Models can Strategically Underperform on Evaluationscited by 145
  14. 2025When Chain of Thought is Necessary, Language Models Struggle to Evade Monitorscited by 88
  15. 2023How to Catch an AI Liar: Lie Detection in Black-Box LLMs by Asking Unrelated Questionscited by 83
  16. 2025Auditing language models for hidden objectivescited by 83
  17. 2024Truth is Universal: Robust Detection of Lies in LLMscited by 79
  18. 2023Still No Lie Detector for Language Models: Probing Empirical and Conceptual Roadblockscited by 78
  19. 2025The MASK Benchmark: Disentangling Honesty From Accuracy in AI Systemscited by 72
  20. 2025Detecting Strategic Deception Using Linear Probescited by 68
  21. 2023Preventing Language Models From Hiding Their Reasoningcited by 65
  22. 2025Detecting High-Stakes Interactions with Activation Probescited by 53
  23. 2024Obfuscated Activations Bypass LLM Latent-Space Defensescited by 45
  24. 2023Honesty Is the Best Policy: Defining and Mitigating AI Deceptioncited by 40
  25. 2024Uncovering Deceptive Tendencies in Language Models: A Simulated Company AI Assistantcited by 38
  26. 2025Evaluating Frontier Models for Stealth and Situational Awarenesscited by 37
  27. 2025Eliciting Secret Knowledge from Language Modelscited by 27
  28. 2025Among Us: A Sandbox for Measuring and Detecting Agentic Deceptioncited by 26
  29. 2023Localizing Lying in Llama: Understanding Instructed Dishonesty on True-False Questions Through Prompting, Probing, and Patchingcited by 24
  30. 2025Believe It or Not: How Deeply do LLMs Believe Implanted Facts?cited by 24
  31. 2025Liars' Bench: Evaluating Lie Detectors for Language Modelscited by 22
  32. 2025MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hackingcited by 19
  33. 2025Towards eliciting latent knowledge from LLMs with mechanistic interpretabilitycited by 19
  34. 2025Benchmarking Deception Probes via Black-to-White Performance Boostscited by 19
  35. 2020An overview of 11 proposals for building safe advanced AIcited by 18
  36. 2025Technological folie `a deux: Feedback Loops Between AI Chatbots and Mental Illnesscited by 15
  37. 2025Preference Learning with Lie Detectors can Induce Honesty or Evasioncited by 9
  38. 2025PROMPTEVALS: A Dataset of Assertions and Guardrails for Custom Production Large Language Model Pipelinescited by 6

50 most cited of 51 All 51 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 13.

13 papers.

Sort:

By area:

  1. 2026Position: Anthropomorphic Misalignment Research Needs Stronger Evidencecited by 7
  2. 2026Linear Probe Accuracy Scales with Model Size and Benefits from Multi-Layer Ensemblingcited by 4
  3. 2026When Role-playing, Do Models Believe What They Say?cited by 4
  4. 2026"Did you lie?" Evaluating Lie Detectors across Model Scale and Belief-Verified Model Organismscited by 4
  5. 2026Instrumental Choices: Measuring the Propensity of LLM Agents to Pursue Instrumental Behaviorscited by 3
  6. 2026Think Before You Lie: How Reasoning Leads to Honestycited by 2
  7. 2026Sycophancy as a Multilingual Alignment Failure: How Safety Degrades Across Languages, Topics, and Modelscited by 1
  8. 2026Probing the Limits of the Lie Detector Approach to LLM Deceptioncited by 0
  9. 2026A Model of Multi-turn Human Persuadability Using Probabilistic Belief Tracingcited by 0
  10. 2026Confidently Deceptive: On the Relationship Between Confidence and Deception in LLMscited by 0
  11. 2026Stress-Testing LLM Lie Detectors: Role-Play Failures and Spurious Correlationscited by 0
  12. 2026Caught in the Act: Probes Effectively Detect Sabotage and Catch Unverbalized Deceptioncited by 0
1 more
  1. 2026The Agent Knew: Internal States Forecast Unauthorized Actions and Stop Them, and the Lies That Followcited by 0