# Unveiling the Lack of LVLM Robustness to Fundamental Visual Variations: Why and Path Forward

arXiv: https://arxiv.org/abs/2504.16727 · HTML: https://mlsanity.com/paper/2504.16727/

## Key facts

- Cited by 7 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 45 of its 69 references that are in the map.
- First posted on arXiv 23 Apr 2025.

## 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, 30 of 45)

- [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)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [The Power of Scale for Parameter-Efficient Prompt Tuning](https://mlsanity.com/paper/2104.08691/) (2021)
- [LLaVA-OneVision: Easy Visual Task Transfer](https://mlsanity.com/paper/2408.03326/) (2024)
- [InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks](https://mlsanity.com/paper/2312.14238/) (2023)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [MMBench: Is Your Multi-modal Model an All-around Player?](https://mlsanity.com/paper/2307.06281/) (2023)
- [MathVista: Evaluating Mathematical Reasoning of Foundation Models in Visual Contexts](https://mlsanity.com/paper/2310.02255/) (2023)
- [Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling](https://mlsanity.com/paper/2412.05271/) (2024)
- [Understanding intermediate layers using linear classifier probes](https://mlsanity.com/paper/1610.01644/) (2016)
- [MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilities](https://mlsanity.com/paper/2308.02490/) (2023)
- [MiniCPM-V: A GPT-4V Level MLLM on Your Phone](https://mlsanity.com/paper/2408.01800/) (2024)
- [ShareGPT4V: Improving Large Multi-Modal Models with Better Captions](https://mlsanity.com/paper/2311.12793/) (2023)
- [Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling](https://mlsanity.com/paper/2501.17811/) (2025)
- [DeepSeek-VL: Towards Real-World Vision-Language Understanding](https://mlsanity.com/paper/2403.05525/) (2024)
- [Understanding the Effective Receptive Field in Deep Convolutional Neural Networks](https://mlsanity.com/paper/1701.04128/) (2017)
- [V*: Guided Visual Search as a Core Mechanism in Multimodal LLMs](https://mlsanity.com/paper/2312.14135/) (2023)
- [Is ChatGPT a Good NLG Evaluator? A Preliminary Study](https://mlsanity.com/paper/2303.04048/) (2023)
- [AppAgent: Multimodal Agents as Smartphone Users](https://mlsanity.com/paper/2312.13771/) (2023)
- [Mobile-Agent: Autonomous Multi-Modal Mobile Device Agent with Visual Perception](https://mlsanity.com/paper/2401.16158/) (2024)
- [MME-RealWorld: Could Your Multimodal LLM Challenge High-Resolution Real-World Scenarios that are Difficult for Humans?](https://mlsanity.com/paper/2408.13257/) (2024)
- [GPT4Tools: Teaching Large Language Model to Use Tools via Self-instruction](https://mlsanity.com/paper/2305.18752/) (2023)
- [JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks](https://mlsanity.com/paper/2404.03027/) (2024)
- [OmniMedVQA: A New Large-Scale Comprehensive Evaluation Benchmark for Medical LVLM](https://mlsanity.com/paper/2402.09181/) (2024)
- [SVIT: Scaling up Visual Instruction Tuning](https://mlsanity.com/paper/2307.04087/) (2023)
- [EmbodiedScan: A Holistic Multi-Modal 3D Perception Suite Towards Embodied AI](https://mlsanity.com/paper/2312.16170/) (2023)

## Built on by (most cited first, 7 of 7)

- [MMBoundary: Advancing MLLM Knowledge Boundary Awareness through Reasoning Step Confidence Calibration](https://mlsanity.com/paper/2505.23224/) (2025)
- [Visually Prompted Benchmarks Are Surprisingly Fragile](https://mlsanity.com/paper/2512.17875/) (2025)
- [CultureCLIP: Empowering CLIP with Cultural Awareness through Synthetic Images and Contextualized Captions](https://mlsanity.com/paper/2507.06210/) (2025)
- ["It's trained by non-disabled people": Evaluating How Image Quality Affects Product Captioning with Vision-Language Models](https://mlsanity.com/paper/2511.08917/) (2025)
- [When Harmful Content Gets Camouflaged: Unveiling Perception Failure of LVLMs with CamHarmTI](https://mlsanity.com/paper/2512.03087/) (2025)
- [The Last Visible Pixel: Probing Fine-Scale Perception in Vision-Language Models](https://mlsanity.com/paper/2606.07861/) (2026)
- [Can't Find Waldo: Evaluating VLMs' Sensitivity to Image Resolution and Detail Level](https://mlsanity.com/paper/2609.31706/) (2026)

## Abstract (arXiv)

Large Vision Language Models (LVLMs) excel in various vision-language tasks. Yet, their robustness to visual variations in position, scale, orientation, and context that objects in natural scenes inevitably exhibit due to changes in viewpoint and environment remains largely underexplored. To bridge this gap, we introduce V2R-Bench, a comprehensive benchmark framework for evaluating Visual Variation Robustness of LVLMs, which encompasses automated evaluation dataset generation and principled metrics for thorough robustness assessment. Through extensive evaluation on 21 LVLMs, we reveal a surprising vulnerability to visual variations, in which even advanced models that excel at complex vision-language tasks significantly underperform on simple tasks such as object recognition. Interestingly, these models exhibit a distinct visual position bias that contradicts theories of effective receptive fields, and demonstrate a human-like visual acuity threshold. To identify the source of these vulnerabilities, we present a systematic framework for component-level analysis, featuring a novel visualization approach for aligned visual features. Results show that these vulnerabilities stem from error accumulation in the pipeline architecture and inadequate multimodal alignment. Complementary experiments with synthetic data further demonstrate that these limitations are fundamentally architectural deficiencies, scoring the need for architectural innovations in future LVLM designs.