# HEDGE: Hallucination Estimation via Dense Geometric Entropy for VQA with Vision-Language Models

arXiv: https://arxiv.org/abs/2511.12693 · HTML: https://mlsanity.com/paper/2511.12693/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Hallucination mitigation in LVLMs, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 51 of its 73 references that are in the map.
- First posted on arXiv 16 Nov 2025.

## Placed in

- Multimodal LLMs / vision-language models / [Hallucination mitigation in LVLMs](https://mlsanity.com/mllm/hallucination-mitigation-in-lvlms/)

## Builds on (most cited first, 30 of 51)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment](https://mlsanity.com/paper/2303.16634/) (2023)
- [DeBERTa: Decoding-enhanced BERT with Disentangled Attention](https://mlsanity.com/paper/2006.03654/) (2020)
- [SimCSE: Simple Contrastive Learning of Sentence Embeddings](https://mlsanity.com/paper/2104.08821/) (2021)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day](https://mlsanity.com/paper/2306.00890/) (2023)
- [When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memories](https://mlsanity.com/paper/2212.10511/) (2022)
- [Object Hallucination in Image Captioning](https://mlsanity.com/paper/1809.02156/) (2018)
- [Albumentations: fast and flexible image augmentations](https://mlsanity.com/paper/1809.06839/) (2018)
- [FreshLLMs: Refreshing Large Language Models with Search Engine Augmentation](https://mlsanity.com/paper/2310.03214/) (2023)
- [AMBER: An LLM-free Multi-dimensional Benchmark for MLLMs Hallucination Evaluation](https://mlsanity.com/paper/2311.07397/) (2023)
- [AlignScore: Evaluating Factual Consistency with a Unified Alignment Function](https://mlsanity.com/paper/2305.16739/) (2023)
- [Understanding Factuality in Abstractive Summarization with FRANK: A Benchmark for Factuality Metrics](https://mlsanity.com/paper/2104.13346/) (2021)
- [Semantic Entropy Probes: Robust and Cheap Hallucination Detection in LLMs](https://mlsanity.com/paper/2406.15927/) (2024)
- [The Effect of Sampling Temperature on Problem Solving in Large Language Models](https://mlsanity.com/paper/2402.05201/) (2024)
- [Lessons from Natural Language Inference in the Clinical Domain](https://mlsanity.com/paper/1808.06752/) (2018)
- [Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lens](https://mlsanity.com/paper/2411.16724/) (2024)
- [A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generation](https://mlsanity.com/paper/2104.08704/) (2021)
- [HalluciDoctor: Mitigating Hallucinatory Toxicity in Visual Instruction Data](https://mlsanity.com/paper/2311.13614/) (2023)
- [FaithDial: A Faithful Benchmark for Information-Seeking Dialogue](https://mlsanity.com/paper/2204.10757/) (2022)
- [Benchmarking Uncertainty Quantification Methods for Large Language Models with LM-Polygraph](https://mlsanity.com/paper/2406.15627/) (2024)
- [SPUQ: Perturbation-Based Uncertainty Quantification for Large Language Models](https://mlsanity.com/paper/2403.02509/) (2024)
- [Polos: Multimodal Metric Learning from Human Feedback for Image Captioning](https://mlsanity.com/paper/2402.18091/) (2024)
- [ALOHa: A New Measure for Hallucination in Captioning Models](https://mlsanity.com/paper/2404.02904/) (2024)

## Built on by (most cited first, 5 of 5)

- [The Geometry of Representational Failures in Vision Language Models](https://mlsanity.com/paper/2602.07025/) (2026)
- [Overconfidence and Calibration in Medical VQA: Empirical Findings and Hallucination-Aware Mitigation](https://mlsanity.com/paper/2604.02543/) (2026)
- [FaithSCAN: Model-Driven Single-Pass Hallucination Detection for Faithful Visual Question Answering](https://mlsanity.com/paper/2601.00269/) (2026)
- [Witness Evidence Portfolios: Single-Prefill Risk Detection for Closed Multimodal Answers](https://mlsanity.com/paper/2607.27667/) (2026)
- [UHP Detection: LVLMs have their Unique Hallucination Pattern in the Consistency Space](https://mlsanity.com/paper/2608.03817/) (2026)

## Abstract (arXiv)

Vision-language models (VLMs) enable open-ended visual question answering but remain prone to hallucinations. We present HEDGE, a unified framework for hallucination detection that combines controlled visual perturbations, semantic clustering, and robust uncertainty metrics. HEDGE integrates sampling, distortion synthesis, clustering (entailment- and embedding-based), and metric computation into a reproducible pipeline applicable across multimodal architectures. Evaluations on VQA-RAD and KvasirVQA-x1 with three representative VLMs (LLaVA-Med, Med-Gemma, Qwen2.5-VL) reveal clear architecture- and prompt-dependent trends. Hallucination detectability is highest for unified-fusion models with dense visual tokenization (Qwen2.5-VL) and lowest for architectures with restricted tokenization (Med-Gemma). Embedding-based clustering often yields stronger separation when applied directly to the generated answers, whereas NLI-based clustering remains advantageous for LLaVA-Med and for longer, sentence-level responses. Across configurations, the VASE metric consistently provides the most robust hallucination signal, especially when paired with embedding clustering and a moderate sampling budget (n ~ 10-15). Prompt design also matters: concise, label-style outputs offer clearer semantic structure than syntactically constrained one-sentence responses. By framing hallucination detection as a geometric robustness problem shaped jointly by sampling scale, prompt structure, model architecture, and clustering strategy, HEDGE provides a principled, compute-aware foundation for evaluating multimodal reliability. The hedge-bench PyPI library enables reproducible and extensible benchmarking, with full code and experimental resources available at https://github.com/Simula/HEDGE .