Hallucination Begins Where Saliency Drops

• URL: http://arxiv.org/abs/2601.20279v1
• Date: Wed, 28 Jan 2026 05:50:52 GMT
• Title: Hallucination Begins Where Saliency Drops
• Authors: Xiaofeng Zhang, Yuanchao Zhu, Chaochen Gu, Xiaosong Yuan, Qiyan Zhao, Jiawei Cao, Feilong Tang, Sinan Fan, Yaomin Shen, Chen Shen, Hao Tang,
• Abstract summary: hallucinations frequently arise when preceding output tokens exhibit low saliency toward the prediction of the next token.<n>We introduce LVLMs-Saliency, a gradient-aware diagnostic framework that quantifies the visual grounding strength of each output token.<n>Our method significantly reduces hallucination rates while preserving fluency and task performance, offering a robust and interpretable solution.
• Score: 18.189047289404325
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Recent studies have examined attention dynamics in large vision-language models (LVLMs) to detect hallucinations. However, existing approaches remain limited in reliably distinguishing hallucinated from factually grounded outputs, as they rely solely on forward-pass attention patterns and neglect gradient-based signals that reveal how token influence propagates through the network. To bridge this gap, we introduce LVLMs-Saliency, a gradient-aware diagnostic framework that quantifies the visual grounding strength of each output token by fusing attention weights with their input gradients. Our analysis uncovers a decisive pattern: hallucinations frequently arise when preceding output tokens exhibit low saliency toward the prediction of the next token, signaling a breakdown in contextual memory retention. Leveraging this insight, we propose a dual-mechanism inference-time framework to mitigate hallucinations: (1) Saliency-Guided Rejection Sampling (SGRS), which dynamically filters candidate tokens during autoregressive decoding by rejecting those whose saliency falls below a context-adaptive threshold, thereby preventing coherence-breaking tokens from entering the output sequence; and (2) Local Coherence Reinforcement (LocoRE), a lightweight, plug-and-play module that strengthens attention from the current token to its most recent predecessors, actively counteracting the contextual forgetting behavior identified by LVLMs-Saliency. Extensive experiments across multiple LVLMs demonstrate that our method significantly reduces hallucination rates while preserving fluency and task performance, offering a robust and interpretable solution for enhancing model reliability. Code is available at: https://github.com/zhangbaijin/LVLMs-Saliency

Related papers

• Look Carefully: Adaptive Visual Reinforcements in Multimodal Large Language Models for Hallucination Mitigation [51.743225614196774]
  Multimodal large language models (MLLMs) have achieved remarkable progress in vision-language reasoning.<n>They remain vulnerable to hallucination, where generated content deviates from visual evidence.<n>Recent vision enhancement methods attempt to address this issue by reinforcing visual tokens during decoding.<n>We propose Adaptive Visual Reinforcement (AIR), a training-free framework for MLLMs.
  arXiv  Detail & Related papers  (2026-02-27T14:18:51Z)
• Context-Aware Decoding for Faithful Vision-Language Generation [5.258492912374723]
  Hallucinations, generating responses inconsistent with the visual input, remain a critical limitation of large vision-language models (LVLMs)<n>We probe the layer-wise generation dynamics that drive hallucinations and propose a training-free mitigation strategy.
  arXiv  Detail & Related papers  (2026-01-09T16:50:57Z)
• FaithSCAN: Model-Driven Single-Pass Hallucination Detection for Faithful Visual Question Answering [14.550872089352943]
  FaithSCAN is a lightweight network that detects hallucinations by exploiting rich internal signals of vision-language models.<n>We extend the LLM-as-a-Judge paradigm to VQA hallucination and propose a low-cost strategy to automatically generate model-dependent supervision signals.<n>In-depth analysis shows hallucinations arise from systematic internal state variations in visual perception, cross-modal reasoning, and language decoding.
  arXiv  Detail & Related papers  (2026-01-01T09:19:39Z)
• Revealing Perception and Generation Dynamics in LVLMs: Mitigating Hallucinations via Validated Dominance Correction [59.801614364841775]
  Large Vision-Language Models (LVLMs) have shown remarkable capabilities, yet hallucinations remain a persistent challenge.<n>This work presents a systematic analysis of the internal evolution of visual perception and token generation in LVLMs.<n>We devise the VDC (d Dominance Correction) strategy, which detects unsupported tokens and replaces them with validated ones to improve output reliability.
  arXiv  Detail & Related papers  (2025-12-21T17:05:42Z)
• ICR Probe: Tracking Hidden State Dynamics for Reliable Hallucination Detection in LLMs [50.18087419133284]
  hallucination detection methods leveraging hidden states predominantly focus on static and isolated representations.<n>We introduce a novel metric, the ICR Score, which quantifies the contribution of modules to the hidden states' update.<n>We propose a hallucination detection method, the ICR Probe, which captures the cross-layer evolution of hidden states.
  arXiv  Detail & Related papers  (2025-07-22T11:44:26Z)
• Grounding Language with Vision: A Conditional Mutual Information Calibrated Decoding Strategy for Reducing Hallucinations in LVLMs [51.93737995405164]
  Large Vision-Language Models (LVLMs) are susceptible to hallucinations.<n>We introduce a novel Conditional Pointwise Mutual Information (C-PMI) calibrated decoding strategy.<n>We show that the proposed method significantly reduces hallucinations in LVLMs while preserving decoding efficiency.
  arXiv  Detail & Related papers  (2025-05-26T08:36:10Z)
• Attention Reallocation: Towards Zero-cost and Controllable Hallucination Mitigation of MLLMs [62.9348974370985]
  We propose attention reallocation (AttnReal) to mitigate hallucinations with nearly zero extra cost.<n>Our approach is motivated by the key observations that, MLLM's unreasonable attention distribution causes features to be dominated by historical output tokens.<n>Based on the observations, AttnReal recycles excessive attention from output tokens and reallocates it to visual tokens, which reduces MLLM's reliance on language priors.
  arXiv  Detail & Related papers  (2025-03-11T11:52:37Z)
• Mitigating Hallucination for Large Vision Language Model by Inter-Modality Correlation Calibration Decoding [66.06337890279839]
  Large vision-language models (LVLMs) have shown remarkable capabilities in visual-language understanding for downstream multi-modal tasks.<n>LVLMs still suffer from generating hallucinations in complex generation tasks, leading to inconsistencies between visual inputs and generated content.<n>We propose an Inter-Modality Correlation Decoding (IMCCD) method to mitigate hallucinations in LVLMs in a training-free manner.
  arXiv  Detail & Related papers  (2025-01-03T17:56:28Z)
• Self-Introspective Decoding: Alleviating Hallucinations for Large Vision-Language Models [30.26685485474035]
  Large Vision-Language Models (LVLMs) have rapidly advanced in recent years.<n>The prevalent issue known as the hallucination' problem has emerged as a significant bottleneck.<n>We propose a simple yet effective method named Self-Introspective Decoding (SID)
  arXiv  Detail & Related papers  (2024-08-04T13:50:17Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.