# Benchmarking Deflection and Hallucination in Large Vision-Language Models

arXiv: https://arxiv.org/abs/2604.12033 · HTML: https://mlsanity.com/paper/2604.12033/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in VLM benchmarks and visual reasoning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 23 of its 35 references that are in the map.
- First posted on arXiv 13 Apr 2026.

## Placed in

- Multimodal LLMs / vision-language models / [VLM benchmarks and visual reasoning](https://mlsanity.com/mllm/vlm-benchmarks-and-visual-reasoning/)

## Builds on (most cited first, 23 of 23)

- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflection](https://mlsanity.com/paper/2310.11511/) (2023)
- [Unsupervised Dense Information Retrieval with Contrastive Learning](https://mlsanity.com/paper/2112.09118/) (2021)
- [OK-VQA: A Visual Question Answering Benchmark Requiring External Knowledge](https://mlsanity.com/paper/1906.00067/) (2019)
- [A-OKVQA: A Benchmark for Visual Question Answering using World Knowledge](https://mlsanity.com/paper/2206.01718/) (2022)
- [VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning](https://mlsanity.com/paper/2504.08837/) (2025)
- [MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries](https://mlsanity.com/paper/2401.15391/) (2024)
- [MuRAG: Multimodal Retrieval-Augmented Generator for Open Question Answering over Images and Text](https://mlsanity.com/paper/2210.02928/) (2022)
- [Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions?](https://mlsanity.com/paper/2302.11713/) (2023)
- [Encyclopedic VQA: Visual questions about detailed properties of fine-grained categories](https://mlsanity.com/paper/2306.09224/) (2023)
- [EVA-CLIP-18B: Scaling CLIP to 18 Billion Parameters](https://mlsanity.com/paper/2402.04252/) (2024)
- [Wiki-LLaVA: Hierarchical Retrieval-Augmented Generation for Multimodal LLMs](https://mlsanity.com/paper/2404.15406/) (2024)
- [Kwai Keye-VL 1.5 Technical Report](https://mlsanity.com/paper/2509.01563/) (2025)
- [EchoSight: Advancing Visual-Language Models with Wiki Knowledge](https://mlsanity.com/paper/2407.12735/) (2024)
- [SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge](https://mlsanity.com/paper/2509.07968/) (2025)
- [Benchmarking Retrieval-Augmented Multimodal Generation for Document Question Answering](https://mlsanity.com/paper/2505.16470/) (2025)
- [HaloQuest: A Visual Hallucination Dataset for Advancing Multimodal Reasoning](https://mlsanity.com/paper/2407.15680/) (2024)
- [LEMMA: Towards LVLM-Enhanced Multimodal Misinformation Detection with External Knowledge Augmentation](https://mlsanity.com/paper/2402.11943/) (2024)
- [SnapNTell: Enhancing Entity-Centric Visual Question Answering with Retrieval Augmented Multimodal LLM](https://mlsanity.com/paper/2403.04735/) (2024)
- [GaRAGe: A Benchmark with Grounding Annotations for RAG Evaluation](https://mlsanity.com/paper/2506.07671/) (2025)
- [Visual-RAG: Benchmarking Text-to-Image Retrieval Augmented Generation for Visual Knowledge Intensive Queries](https://mlsanity.com/paper/2502.16636/) (2025)
- [SK-VQA: Synthetic Knowledge Generation at Scale for Training Context-Augmented Multimodal LLMs](https://mlsanity.com/paper/2406.19593/) (2024)

## Built on by (most cited first, 1 of 1)

- [MedVIGIL: Evaluating Trustworthy Medical VLMs Under Broken Visual Evidence](https://mlsanity.com/paper/2605.07919/) (2026)

## Abstract (arXiv)

Large Vision-Language Models (LVLMs) increasingly rely on retrieval to answer knowledge-intensive multimodal questions. Existing benchmarks overlook conflicts between visual and textual evidence and the importance of generating deflections (e.g., Sorry, I cannot answer...) when retrieved knowledge is incomplete. These benchmarks also suffer from rapid obsolescence, as growing LVLM training sets allow models to answer many questions without retrieval. We address these gaps with three contributions. First, we propose a dynamic data curation pipeline that preserves benchmark difficulty over time by filtering for genuinely retrieval-dependent samples. Second, we introduce VLM-DeflectionBench, a benchmark of 2,775 samples spanning diverse multimodal retrieval settings, designed to probe model behaviour under conflicting or insufficient evidence. Third, we define a fine-grained evaluation protocol with four scenarios that disentangle parametric memorization from retrieval robustness. Experiments across 20 state-of-the-art LVLMs indicate that models usually fail to deflect in the presence of noisy or misleading evidence. Our results highlight the need to evaluate not only what models know, but how they behave when they do not, and serve as a reusable and extensible benchmark for reliable KB-VQA evaluation. All resources will be publicly available upon publication.