# Beyond Cross-Modal Alignment: Measuring and Leveraging Modality Gap in Vision-Language Models

arXiv: https://arxiv.org/abs/2502.14888 · HTML: https://mlsanity.com/paper/2502.14888/

## Key facts

- Cited by 4 papers in the map (data through 14 Oct 2026).
- Placed in MLLM training and visual instruction tuning, one of 16 directions in Multimodal LLMs / vision-language models.
- Builds on 26 of its 46 references that are in the map.
- First posted on arXiv 16 Feb 2025.

## Placed in

- Multimodal LLMs / vision-language models / [MLLM training and visual instruction tuning](https://mlsanity.com/mllm/mllm-training-and-visual-instruction-tuning/)

## Builds on (most cited first, 26 of 26)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [ViLBERT: Pretraining Task-Agnostic Visiolinguistic Representations for Vision-and-Language Tasks](https://mlsanity.com/paper/1908.02265/) (2019)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [A Survey on LLM-as-a-Judge](https://mlsanity.com/paper/2411.15594/) (2024)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [ViLT: Vision-and-Language Transformer Without Convolution or Region Supervision](https://mlsanity.com/paper/2102.03334/) (2021)
- [Aligning Large Multimodal Models with Factually Augmented RLHF](https://mlsanity.com/paper/2309.14525/) (2023)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Supervision Exists Everywhere: A Data Efficient Contrastive Language-Image Pre-training Paradigm](https://mlsanity.com/paper/2110.05208/) (2021)
- [Finding Neurons in a Haystack: Case Studies with Sparse Probing](https://mlsanity.com/paper/2305.01610/) (2023)
- [Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models](https://mlsanity.com/paper/2307.14539/) (2023)
- [Generating Images with Multimodal Language Models](https://mlsanity.com/paper/2305.17216/) (2023)
- [Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations](https://mlsanity.com/paper/2410.02762/) (2024)
- [On the Robustness of Large Multimodal Models Against Image Adversarial Attacks](https://mlsanity.com/paper/2312.03777/) (2023)
- [Two Effects, One Trigger: On the Modality Gap, Object Bias, and Information Imbalance in Contrastive Vision-Language Models](https://mlsanity.com/paper/2404.07983/) (2024)
- [SPAE: Semantic Pyramid AutoEncoder for Multimodal Generation with Frozen LLMs](https://mlsanity.com/paper/2306.17842/) (2023)
- [MM-SHAP: A Performance-agnostic Metric for Measuring Multimodal Contributions in Vision and Language Models & Tasks](https://mlsanity.com/paper/2212.08158/) (2022)
- [Do Vision & Language Decoders use Images and Text equally? How Self-consistent are their Explanations?](https://mlsanity.com/paper/2404.18624/) (2024)
- [Beyond Logit Lens: Contextual Embeddings for Robust Hallucination Detection & Grounding in VLMs](https://mlsanity.com/paper/2411.19187/) (2024)
- [Multi-modal Deep Learning](https://mlsanity.com/paper/2403.03385/) (2024)
- [MultiModN- Multimodal, Multi-Task, Interpretable Modular Networks](https://mlsanity.com/paper/2309.14118/) (2023)
- [UnifiedMLLM: Enabling Unified Representation for Multi-modal Multi-tasks With Large Language Model](https://mlsanity.com/paper/2408.02503/) (2024)
- [Encourage or Inhibit Monosemanticity? Revisit Monosemanticity from a Feature Decorrelation Perspective](https://mlsanity.com/paper/2406.17969/) (2024)
- [Dissecting Multimodality in VideoQA Transformer Models by Impairing Modality Fusion](https://mlsanity.com/paper/2306.08889/) (2023)

## Built on by (most cited first, 4 of 4)

- [Interpretable Debiasing of Vision-Language Models for Social Fairness](https://mlsanity.com/paper/2602.24014/) (2026)
- [Decomposing multimodal embedding spaces with group-sparse autoencoders](https://mlsanity.com/paper/2601.20028/) (2026)
- [Prior Directions: Why GUI Grounding Gets Locked in the Past](https://mlsanity.com/paper/2607.26913/) (2026)
- [Omni-Modal LLMs Know More Than They Say](https://mlsanity.com/paper/2610.16135/) (2026)

## Abstract (arXiv)

The success of vision-language models is primarily attributed to effective alignment across modalities such as vision and language. However, modality gaps persist in existing alignment algorithms and appear necessary for human perception as evidenced by modality-specific phenomena like visual texture and linguistic tone. These observations motivate us to computationally measure and leverage modality gaps to improve downstream tasks. We first introduce the Modality Dominance Score (MDS), which attributes multimodal features to specific modalities by categorizing them into three classes: vision-dominant features, language-dominant features, and cross-modal features. We then propose automatic interpretability metrics to evaluate these modality-specific features in a scalable manner. Finally, we demonstrate that the training-free model editing enhances multiple downstream tasks, including mitigating bias in gender classification, generating cross-modal adversarial examples, and enabling modality-specific control in text-to-image generation. Combined with task-agnostic interpretability tools, our work offers insights for systematic analysis and lightweight editing of multimodal models.