mlsanityData to 14 Oct 2026

Privileged Self-Access Matters for Introspection in AI

arXiv 2508.14802 ↗First posted 20 Aug 2025cs.AI cs.CLcited by 22

Not on the map. Why?

Abstract (arXiv)

Whether AI models can introspect is an increasingly important practical question. But there is no consensus on how introspection is to be defined. Beginning from a recently proposed ''lightweight'' definition, we argue instead for a thicker one. According to our proposal, introspection in AI is any process which yields information about internal states through a process more reliable than one with equal or lower computational cost available to a third party. Using experiments where LLMs reason about their internal temperature parameters, we show they can appear to have lightweight introspection while failing to meaningfully introspect per our proposed definition.

Key facts. Cited by 22 papers in the map (data through 14 Oct 2026). Builds on 6 of its 12 references that are in the map. First posted on arXiv 20 Aug 2025.

Builds on

6 of 12 references.

References · 6

  1. 2024GPT-4o System Cardcited by 4,588
  2. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  3. 2024Looking Inward: Language Models Can Learn About Themselves by Introspectioncited by 113
  4. 2025Tell me about yourself: LLMs are aware of their learned behaviorscited by 91
  5. 2025Does It Make Sense to Speak of Introspection in Large Language Models?cited by 25
  6. 2024Large Language Models Cannot Explain Themselvescited by 11

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 5; 2026: 17.

22 papers.

Sort:

By area:

  1. 2026Verbalizable Representations Form a Global Workspace in Language Modelscited by 75
  2. 2025Training Language Models to Explain Their Own Computationscited by 30
  3. 2026Latent Introspection: Models Can Detect Prior Concept Injectionscited by 14
  4. 2025Evidence for Limited Metacognition in LLMscited by 12
  5. 2026Can LLMs Introspect? A Reality Checkcited by 12
  6. 2025Steering Awareness: Detecting Activation Steering from Withincited by 9
  7. 2026Emergent Introspection in AI is Content-Agnosticcited by 7
  8. 2025Generalized Correctness Models: Learning Calibrated and Model-Agnostic Correctness Predictors from Historical Patternscited by 5
  9. 2026Hallucinations Undermine Trust; Metacognition is a Way Forwardcited by 5
  10. 2026Introspective Coupling: Self-Explanation Training Tracks Behavioral Change Despite Fixed Supervisioncited by 4
  11. 2025Metacognitive Sensitivity for Test-Time Dynamic Model Selectioncited by 2
  12. 2026Self-Reference in Large Language Models: The Introspection Threshold for Recursive Self-Improvementcited by 2
10 more
  1. 2026When to Call an Apple Red: Humans Follow Introspective Rules, VLMs Don'tcited by 1
  2. 2026Beyond Confidence: Rethinking Self-Assessments for Performance Prediction in LLMscited by 1
  3. 2026Metacognition in LLMs: Foundations, Progress, and Opportunitiescited by 1
  4. 2026Phase Transitions in Driven Informational Systems: A Two-Field Perspective on Learning Theory and Non-Equilibrium Chemistrycited by 0
  5. 2026Operational Proto-Introspection in Looped Language Models: Process-Quality Taps, Executable Branching, and the Readout-Control Boundarycited by 0
  6. 2026Asymmetric Communication: Large Language Models and Language Gamescited by 0
  7. 2026Open-Weight Masked Introspection: Measuring What Language Models Can Report About Their Own Computationcited by 0
  8. 2026In-Context Neurofeedback: Can LLMs Control Their Internal Representations through Privileged Access?cited by 0
  9. 2026Nudgeability: Reasoning Models Follow Confidence Signals Without Tracking Their Own Competencecited by 0
  10. 2026A Sense of Proprioception in Language Modelscited by 0