Pith. sign in

REVIEW 10 cited by

VL-Uncertainty: Detecting Hallucination in Large Vision-Language Model via Uncertainty Estimation

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2411.11919 v2 pith:GMVDS7E5 submitted 2024-11-18 cs.CV

Ruiyang Zhang , Hu Zhang , Zhedong Zheng This is my paper
classification cs.CV
keywords lvlmsuncertaintyvl-uncertaintydetectinghallucinationlvlmresponsessemantically
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Given the higher information load processed by large vision-language models (LVLMs) compared to single-modal LLMs, detecting LVLM hallucinations requires more human and time expense, and thus rise a wider safety concerns. In this paper, we introduce VL-Uncertainty, the first uncertainty-based framework for detecting hallucinations in LVLMs. Different from most existing methods that require ground-truth or pseudo annotations, VL-Uncertainty utilizes uncertainty as an intrinsic metric. We measure uncertainty by analyzing the prediction variance across semantically equivalent but perturbed prompts, including visual and textual data. When LVLMs are highly confident, they provide consistent responses to semantically equivalent queries. However, when uncertain, the responses of the target LVLM become more random. Considering semantically similar answers with different wordings, we cluster LVLM responses based on their semantic content and then calculate the cluster distribution entropy as the uncertainty measure to detect hallucination. Our extensive experiments on 10 LVLMs across four benchmarks, covering both free-form and multi-choice tasks, show that VL-Uncertainty significantly outperforms strong baseline methods in hallucination detection.

Discussion (0). Sign in to comment.

Forward citations

Cited by 10 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. It's the Decoding Format, Not the Perturbation: Auditing Consistency-Based Selection for Vision-Language Test-Time Scaling

    cs.CV 2026-08 conditional novelty 7.0 of 10

    Perturbation-based selection does not beat a format-matched control that spends the same short-answer budget on the unperturbed image, so reported gains against CoT-only majority voting are mostly a decoding-format effect.

  2. HalluScope: Fine-grained Hallucination Diagnosis for Multimodal Large Language Models

    cs.CV 2026-07 conditional novelty 6.0 of 10

    HalluScope couples span-level hallucination detection, 12-way type classification, and explanation generation in one model, and shows the resulting feedback reduces hallucinations in two MLLMs.

  3. When Does Visual Token Pruning Improve Calibration? The Role of Evidence Coverage in MLLMs

    cs.CV 2026-04 unverdicted novelty 6.0 of 10

    Coverage-based visual token pruning at K=128 lowers LLaVA-1.5's POPE ECE from 0.041 to 0.016 without an accuracy drop, and the token selection rule, not the budget, drives calibration quality.

  4. Deterministic Hallucination Detection in Medical VQA via Confidence-Evidence Bayesian Gain

    cs.AI 2026-03 conditional novelty 6.0 of 10

    A fixed, sampling-free score — per-token log-probability variance times (1 + average |image-vs-text probability shift|) — detects medical-VQA hallucinations better than semantic-entropy baselines in 13 of 16 settings.

  5. MIRAGE: Assessing Hallucination in Multimodal Reasoning Chains of MLLM

    cs.CV 2025-05 conditional novelty 6.0 of 10

    MIRAGE is a benchmark that separates reasoning hallucinations from perception errors in multimodal LLMs, and Logos is a curriculum reinforcement fine-tuning method that reduces logical hallucinations.

  6. Analysis of Image-and-Text Uncertainty Propagation in Multimodal Large Language Models with Cardiac MR-Based Applications

    cs.CV 2025-07 conditional novelty 5.0 of 10

    A linear uncertainty-propagation model fitted on cardiac MRI plus health-record text is shown to transfer across prediction tasks and data distributions, enabling cheaper uncertainty estimates.

  7. Uncertainty-o: One Model-agnostic Framework for Unveiling Uncertainty in Large Multimodal Models

    cs.CV 2025-06 conditional novelty 5.0 of 10

    Uncertainty-o estimates uncertainty in large multimodal models by perturbing prompts and computing entropy over semantically clustered answers, improving hallucination detection across five modalities.

  8. LEC: Linear Expectation Constraints for Selection-Conditioned Risk Control in Selective Prediction and Routing Systems

    cs.AI 2025-12 reject novelty 4.0 of 10

    LEC proposes a +1-corrected threshold for FDR control in selective prediction and routing, but the finite-sample guarantee rests on a false exchangeability identity and is not valid for arbitrary exchangeable data.

  9. Mitigating Behavioral Hallucination in Multimodal Large Language Models for Sequential Images

    cs.AI 2025-06 reject novelty 4.0 of 10

    SHE lowers behavioral hallucination scores by about 10 percent by detecting low visual-textual similarity and projecting out the hallucinated direction in embedding space.

  10. Empowering Multimodal LLMs with External Tools: A Comprehensive Survey

    cs.CV 2025-08 unverdicted novelty 2.0 of 10

    A survey paper maps how external tools are used to augment multimodal large language models across data, tasks, evaluation, and future directions.

Pith tools