mlsanityData to 14 Oct 2026

Anosognosia in LLMs: Probing Self-Awareness of Quantized Computational Substrate

arXiv 2610.06174 ↗First posted 5 Oct 2026cs.LG cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

Can LLMs recognize degradation in their own computational substrate? Inspired by anosognosia, a neurological condition in which patients fail to recognize impairments in their own abilities, we investigate whether LLMs can recognize degradation in their computational substrate induced by quantization. We first show that existing models fail to self-report their quantization state, even when provided with their own generated text as an external cue. Linear probing reveals that, while generated text carries almost no trace of quantization, internal representations contain clear, method-specific fingerprints. Through training, models learn to identify severely degraded outputs such as those of 4-bit models by comparison, yet still fail to do so from a single output. A shared LoRA trained jointly across quantization levels succeeded in reading out internal fingerprints, but fails on unseen quantization methods, merely mapping method-specific fingerprints to labels. Whereas external self-observation can restore awareness in some cases of human anosognosia, our results suggest that the more promising route to enabling such awareness in LLMs may lie in their internal representations. Our results highlight fundamental limits of generalizability to LLM self-monitoring.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 36 of its 52 references that are in the map. First posted on arXiv 5 Oct 2026.

Builds on

36 of 52 references.

References · 36

  1. 2022Training language models to follow instructions with human feedbackcited by 15,971
  2. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  3. 2023Let's Verify Step by Stepcited by 3,789
  4. 2017Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layercited by 3,509
  5. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,975
  6. 2016Understanding intermediate layers using linear classifier probescited by 1,353
  7. 2018Don't Give Me the Details, Just the Summary! Topic-Aware Convolutional Neural Networks for Extreme Summarizationcited by 1,317
  8. 2023AWQ: Activation-aware Weight Quantization for LLM Compression and Accelerationcited by 1,298
  9. 2018Hierarchical Neural Story Generationcited by 1,193
  10. 2019ELI5: Long Form Question Answeringcited by 575
  11. 2023How is ChatGPT's behavior changing over time?cited by 270
  12. 2021Surface Form Competition: Why the Highest Probability Answer Isn't Always Rightcited by 263
24 more
  1. 2024The Unreasonable Ineffectiveness of the Deeper Layerscited by 226
  2. 2022Extracting Latent Steering Vectors from Pretrained Language Modelscited by 225
  3. 2024AI Sandbagging: Language Models can Strategically Underperform on Evaluationscited by 145
  4. 2023Taken out of context: On measuring situational awareness in LLMscited by 143
  5. 2025Large Language Models Often Know When They Are Being Evaluatedcited by 128
  6. 2024Looking Inward: Language Models Can Learn About Themselves by Introspectioncited by 113
  7. 2025Tell me about yourself: LLMs are aware of their learned behaviorscited by 91
  8. 2026Emergent Introspective Awareness in Large Language Modelscited by 82
  9. 2024Uncovering Safety Risks of Large Language Models through Concept Activation Vectorcited by 67
  10. 2024Exploiting LLM Quantizationcited by 65
  11. 2023A Setwise Approach for Effective and Highly Efficient Zero-shot Ranking with Large Language Modelscited by 55
  12. 2025Are You Getting What You Pay For? Auditing Model Substitution in LLM APIscited by 38
  13. 2025Steering Evaluation-Aware Language Models to Act Like They Are Deployedcited by 35
  14. 2025Language Models Fail to Introspect About Their Knowledge of Languagecited by 30
  15. 2023From Text to Source: Results in Detecting Large Language Model-Generated Contentcited by 26
  16. 2025Auditing Black-Box LLM APIs with a Rank-Based Uniformity Testcited by 18
  17. 2024Look at the Text: Instruction-Tuned Language Models are More Robust Multiple Choice Selectors than You Thinkcited by 14
  18. 2026Can LLMs Introspect? A Reality Checkcited by 12
  19. 2025Learning to Interpret Weight Differences in Language Modelscited by 11
  20. 2026Introspection Adapters: Training LLMs to Report Their Learned Behaviorscited by 11
  21. 2025Do Large Language Models Know What They Are Capable Of?cited by 10
  22. 2026Mechanisms of Introspective Awarenesscited by 7
  23. 2026Steering at the Source: Style Modulation Heads for Robust Persona Controlcited by 5
  24. 2025On-the-Fly Adaptation to Quantization: Configuration-Aware LoRA for Efficient Fine-Tuning of Quantized LLMscited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2025Interpreting the Effects of Quantization on LLMs
  2. 2024LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinations
  3. 2025Minimal and Mechanistic Conditions for Behavioral Self-Awareness in LLMs
  4. 2025Can LLMs Predict Their Own Failures? Self-Awareness via Internal Circuits
  5. 2026Causally Grounded Mechanistic Interpretability for LLMs with Faithful Natural-Language Explanations
  6. 2026Understanding Large Language Models
  7. 2024MM-SAP: A Comprehensive Benchmark for Assessing Self-Awareness of Multimodal Large Language Models in Perception
  8. 2023Do Emergent Abilities Exist in Quantized Large Language Models: An Empirical Study