Saved in:
Bibliographic Details
Main Authors: Zhang, Dong-Xiao, Lou, Hu, Zhang, Jun-Jie, Zhu, Jun, Meng, Deyu
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.19562
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908916712996864
author Zhang, Dong-Xiao
Lou, Hu
Zhang, Jun-Jie
Zhu, Jun
Meng, Deyu
author_facet Zhang, Dong-Xiao
Lou, Hu
Zhang, Jun-Jie
Zhu, Jun
Meng, Deyu
contents Adversarial vulnerability in vision and hallucination in large language models are conventionally viewed as separate problems, each addressed with modality-specific patches. This study first reveals that they share a common geometric origin: the input and its loss gradient are conjugate observables subject to an irreducible uncertainty bound. Formalizing a Neural Uncertainty Principle (NUP) under a loss-induced state, we find that in near-bound regimes, further compression must be accompanied by increased sensitivity dispersion (adversarial fragility), while weak prompt-gradient coupling leaves generation under-constrained (hallucination). Crucially, this bound is modulated by an input-gradient correlation channel, captured by a specifically designed single-backward probe. In vision, masking highly coupled components improves robustness without costly adversarial training; in language, the same prefill-stage probe detects hallucination risk before generating any answer tokens. NUP thus turns two seemingly separate failure taxonomies into a shared uncertainty-budget view and provides a principled lens for reliability analysis. Guided by this NUP theory, we propose ConjMask (masking high-contribution input components) and LogitReg (logit-side regularization) to improve robustness without adversarial training, and use the probe as a decoding-free risk signal for LLMs, enabling hallucination detection and prompt selection. NUP thus provides a unified, practical framework for diagnosing and mitigating boundary anomalies across perception and generation tasks.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19562
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Neural Uncertainty Principle: A Unified View of Adversarial Fragility and LLM Hallucination
Zhang, Dong-Xiao
Lou, Hu
Zhang, Jun-Jie
Zhu, Jun
Meng, Deyu
Machine Learning
Information Theory
Computational Physics
Adversarial vulnerability in vision and hallucination in large language models are conventionally viewed as separate problems, each addressed with modality-specific patches. This study first reveals that they share a common geometric origin: the input and its loss gradient are conjugate observables subject to an irreducible uncertainty bound. Formalizing a Neural Uncertainty Principle (NUP) under a loss-induced state, we find that in near-bound regimes, further compression must be accompanied by increased sensitivity dispersion (adversarial fragility), while weak prompt-gradient coupling leaves generation under-constrained (hallucination). Crucially, this bound is modulated by an input-gradient correlation channel, captured by a specifically designed single-backward probe. In vision, masking highly coupled components improves robustness without costly adversarial training; in language, the same prefill-stage probe detects hallucination risk before generating any answer tokens. NUP thus turns two seemingly separate failure taxonomies into a shared uncertainty-budget view and provides a principled lens for reliability analysis. Guided by this NUP theory, we propose ConjMask (masking high-contribution input components) and LogitReg (logit-side regularization) to improve robustness without adversarial training, and use the probe as a decoding-free risk signal for LLMs, enabling hallucination detection and prompt selection. NUP thus provides a unified, practical framework for diagnosing and mitigating boundary anomalies across perception and generation tasks.
title Neural Uncertainty Principle: A Unified View of Adversarial Fragility and LLM Hallucination
topic Machine Learning
Information Theory
Computational Physics
url https://arxiv.org/abs/2603.19562