mlsanityData to 14 Oct 2026

Can Linear Probes Measure LLM Uncertainty?

arXiv 2510.04108 ↗First posted 5 Oct 2025cs.LG cs.NA math.NAcited by 3

On the map: Mechanistic interpretability → SAE features and their applications

Abstract (arXiv)

Effective Uncertainty Quantification (UQ) represents a key aspect for reliable deployment of Large Language Models (LLMs) in automated decision-making and beyond. Yet, for LLM generation with multiple choice structure, the state-of-the-art in UQ is still dominated by the naive baseline given by the maximum softmax score. To address this shortcoming, we demonstrate that taking a principled approach via Bayesian statistics leads to improved performance despite leveraging the simplest possible model, namely linear regression. More precisely, we propose to train multiple Bayesian linear models, each predicting the output of a layer given the output of the previous one. Based on the obtained layer-level posterior distributions, we infer the global uncertainty level of the LLM by identifying a sparse combination of distributional features, leading to an efficient UQ scheme. Numerical experiments on various LLMs show consistent improvement over state-of-the-art baselines.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Placed in SAE features and their applications, one of 13 directions in Mechanistic interpretability. Builds on 17 of its 31 references that are in the map. First posted on arXiv 5 Oct 2025.

Builds on

17 of 31 references.

References · 17

  1. 2024The Llama 3 Herd of Modelscited by 14,715
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  4. 2017What Uncertainties Do We Need in Bayesian Deep Learning for Computer Vision?cited by 2,352
  5. 2022Language Models (Mostly) Know What They Knowcited by 1,856
  6. 2023Semantic Uncertainty: Linguistic Invariances for Uncertainty Estimation in Natural Language Generationcited by 819
  7. 2023Eliciting Latent Predictions from Transformers with the Tuned Lenscited by 623
  8. 2024INSIDE: LLMs' Internal States Retain the Power of Hallucination Detectioncited by 323
  9. 2023Generating with Confidence: Uncertainty Quantification for Black-box Large Language Modelscited by 293
  10. 2023Shifting Attention to Relevance: Towards the Predictive Uncertainty Quantification of Free-Form Large Language Modelscited by 150
  11. 2022Analyzing Transformers in Embedding Spacecited by 132
  12. 2024Benchmarking Uncertainty Quantification Methods for Large Language Models with LM-Polygraphcited by 105
5 more
  1. 2024Unsupervised Real-Time Hallucination Detection based on the Internal States of Large Language Modelscited by 86
  2. 2023Overthinking the Truth: Understanding how Language Models Process False Demonstrationscited by 81
  3. 2019On Robustness of Principal Component Regressioncited by 28
  4. 2023ProGAP: Progressive Graph Neural Networks with Differential Privacy Guaranteescited by 12
  5. 2025Machine Unlearning Meets Adversarial Robustness via Constrained Interventions on LLMscited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 3.

3 papers.

  1. 2026The Origins of Stochasticity: Comprehensive Investigations on Uncertainty Quantification for Large Language Modelscited by 2
  2. 2026Reading Calibrated Uncertainty from Language Model Trajectoriescited by 1
  3. 2026Low-Resource Preference Adaptation of LLMs via Activation-Based Label Propagationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Uncertainty Quantification and Confidence Calibration in Large Language Models: A Survey
  2. 2025Simple Yet Effective: An Information-Theoretic Approach to Multi-LLM Uncertainty Quantification
  3. 2026Uncertainty Quantification for LLM Function-Calling
  4. 2025Revisiting Uncertainty Estimation and Calibration of Large Language Models
  5. 2026Between the Layers Lies the Truth: Uncertainty Estimation in LLMs Using Intra-Layer Local Information Scores
  6. 2024Functional-level Uncertainty Quantification for Calibrated Fine-tuning on LLMs
  7. 2025CoT-UQ: Improving Response-wise Uncertainty Quantification in LLMs with Chain-of-Thought
  8. 2025Calibrating Uncertainty Quantification of Multi-Modal LLMs using Grounding