# How Visual Representations Map to Language Feature Space in Multimodal LLMs

arXiv: https://arxiv.org/abs/2506.11976 · HTML: https://mlsanity.com/paper/2506.11976/

## Key facts

- Cited by 25 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 19 of its 26 references that are in the map.
- First posted on arXiv 13 Jun 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)
- Multimodal LLMs / vision-language models / [MLLM training and visual instruction tuning](https://mlsanity.com/mllm/mllm-training-and-visual-instruction-tuning/)
- Mechanistic interpretability / [Interpreting vision and multimodal models](https://mlsanity.com/mechinterp/interpreting-vision-and-multimodal-models/)

## Builds on (most cited first, 19 of 19)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [Gemma 2: Improving Open Language Models at a Practical Size](https://mlsanity.com/paper/2408.00118/) (2024)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Models](https://mlsanity.com/paper/2306.13394/) (2023)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [VILA: On Pre-training for Visual Language Models](https://mlsanity.com/paper/2312.07533/) (2023)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Interpreting CLIP's Image Representation via Text-Based Decomposition](https://mlsanity.com/paper/2310.05916/) (2023)
- [Open Problems in Mechanistic Interpretability](https://mlsanity.com/paper/2501.16496/) (2025)
- [Linearly Mapping from Image to Text Space](https://mlsanity.com/paper/2209.15162/) (2022)
- [Towards Interpreting Visual Information Processing in Vision-Language Models](https://mlsanity.com/paper/2410.07149/) (2024)
- [Interpreting and Editing Vision-Language Representations to Mitigate Hallucinations](https://mlsanity.com/paper/2410.02762/) (2024)
- [Towards Vision-Language Mechanistic Interpretability: A Causal Tracing Tool for BLIP](https://mlsanity.com/paper/2308.14179/) (2023)
- [Multimodal Neurons in Pretrained Text-Only Transformers](https://mlsanity.com/paper/2308.01544/) (2023)
- [Interpreting and Controlling Vision Foundation Models via Text Explanations](https://mlsanity.com/paper/2310.10591/) (2023)
- [Bridging Vision and Language Spaces with Assignment Prediction](https://mlsanity.com/paper/2404.09632/) (2024)

## Built on by (most cited first, 25 of 25)

- [Visual Representation Alignment for Multimodal Large Language Models](https://mlsanity.com/paper/2509.07979/) (2025)
- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [LatentLens: Revealing Highly Interpretable Visual Tokens in LLMs](https://mlsanity.com/paper/2602.00462/) (2026)
- [Understanding Language Prior of LVLMs by Contrasting Chain-of-Embedding](https://mlsanity.com/paper/2509.23050/) (2025)
- [Rethinking Visual Token Reduction in LVLMs Under Cross-Modal Misalignment](https://mlsanity.com/paper/2506.22283/) (2025)
- [Too Late to Recall: Explaining the Two-Hop Problem in Multimodal Knowledge Retrieval](https://mlsanity.com/paper/2512.03276/) (2025)
- [What Do Visual Tokens Really Encode? Uncovering Sparsity and Redundancy in Multimodal Large Language Models](https://mlsanity.com/paper/2603.00510/) (2026)
- [Interpretable and Steerable Concept Bottleneck Sparse Autoencoders](https://mlsanity.com/paper/2512.10805/) (2025)
- [Pathways of Visual Information Flow in Vision-Language Models](https://mlsanity.com/paper/2607.03358/) (2026)
- [Representations of Text and Images Align From Layer One](https://mlsanity.com/paper/2601.08017/) (2026)
- [Unleashing the Intrinsic Visual Representation Capability of Multimodal Large Language Models](https://mlsanity.com/paper/2512.06281/) (2025)
- [Language Models Can Explain Visual Features via Steering](https://mlsanity.com/paper/2603.22593/) (2026)
- [Interpretable Open-Vocabulary Referring Object Detection with Reverse Contrast Attention](https://mlsanity.com/paper/2507.19891/) (2025)
- [Towards Understanding Multimodal Fine-Tuning: Spatial Features](https://mlsanity.com/paper/2602.08713/) (2026)
- [Interpretability Transfer from Language to Vision via Sparse Autoencoders](https://mlsanity.com/paper/2605.24946/) (2026)
- [Representation Alignment Rests on Linear Structure](https://mlsanity.com/paper/2605.28870/) (2026)
- [Slow to See, Slow to Suppress: Understanding the Effects of Modality in Context-Memory Conflicts](https://mlsanity.com/paper/2609.00293/) (2026)
- [Understanding Temporal Logic Consistency in Video-Language Models through Cross-Modal Attention Discriminability](https://mlsanity.com/paper/2510.08138/) (2025)
- [Visual Instruction Tuning Aligns Modalities through Abstraction](https://mlsanity.com/paper/2606.03871/) (2026)
- [Multimodal Model Diffing for Feature Discovery and Control](https://mlsanity.com/paper/2608.09928/) (2026)
- [Why Vision Fails as a Universal Bridge: Rectifying Modality Asynchrony in Multilingual MLLMs](https://mlsanity.com/paper/2608.15085/) (2026)
- [Using OCR Heads to Verbalize Image Semantics](https://mlsanity.com/paper/2609.18823/) (2026)
- [Virtual Encoders in Multimodal Transformers](https://mlsanity.com/paper/2609.26513/) (2026)
- [Resolving State-Representation Mismatch: State-Space Visual Reasoning for Open-Loop VLA Planning](https://mlsanity.com/paper/2609.33412/) (2026)
- [Geometric Similarity in VLM Low-Level Vision Representations](https://mlsanity.com/paper/2610.00848/) (2026)

## Abstract (arXiv)

Effective multimodal reasoning depends on the alignment of visual and linguistic representations, yet the mechanisms by which vision-language models (VLMs) achieve this alignment remain poorly understood. Following the LiMBeR framework, we deliberately maintain a frozen large language model (LLM) and a frozen vision transformer (ViT), connected solely by training a linear adapter during visual instruction tuning. By keeping the language model frozen, we ensure it maintains its original language representations without adaptation to visual data. Consequently, the linear adapter must map visual features directly into the LLM's existing representational space rather than allowing the language model to develop specialized visual understanding through fine-tuning. Our experimental design uniquely enables the use of pre-trained sparse autoencoders (SAEs) of the LLM as analytical probes. These SAEs remain perfectly aligned with the unchanged language model and serve as a snapshot of the learned language feature-representations. Through systematic analysis of SAE reconstruction error, sparsity patterns, and feature SAE descriptions, we reveal the layer-wise progression through which visual representations gradually align with language feature representations, converging in middle-to-later layers. This suggests a fundamental misalignment between ViT outputs and early LLM layers, raising important questions about whether current adapter-based architectures optimally facilitate cross-modal representation learning.