# A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models

arXiv: https://arxiv.org/abs/2502.17516 · HTML: https://mlsanity.com/paper/2502.17516/

## Key facts

- Cited by 34 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 136 of its 180 references that are in the map.
- First posted on arXiv 22 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/)
- Mechanistic interpretability / [Interpreting vision and multimodal models](https://mlsanity.com/mechinterp/interpreting-vision-and-multimodal-models/)

## Builds on (most cited first, 30 of 136)

- [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)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Hierarchical Text-Conditional Image Generation with CLIP Latents](https://mlsanity.com/paper/2204.06125/) (2022)
- [A Unified Approach to Interpreting Model Predictions](https://mlsanity.com/paper/1705.07874/) (2017)
- ["Why Should I Trust You?": Explaining the Predictions of Any Classifier](https://mlsanity.com/paper/1602.04938/) (2016)
- [Mamba: Linear-Time Sequence Modeling with Selective State Spaces](https://mlsanity.com/paper/2312.00752/) (2023)
- [SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis](https://mlsanity.com/paper/2307.01952/) (2023)
- [Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervision](https://mlsanity.com/paper/2102.05918/) (2021)
- [Sigmoid Loss for Language Image Pre-Training](https://mlsanity.com/paper/2303.15343/) (2023)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Prompt-to-Prompt Image Editing with Cross Attention Control](https://mlsanity.com/paper/2208.01626/) (2022)
- [Understanding Black-box Predictions via Influence Functions](https://mlsanity.com/paper/1703.04730/) (2017)
- [Federated Learning: Challenges, Methods, and Future Directions](https://mlsanity.com/paper/1908.07873/) (2019)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Transformer Feed-Forward Layers Are Key-Value Memories](https://mlsanity.com/paper/2012.14913/) (2020)
- [Representation Engineering: A Top-Down Approach to AI Transparency](https://mlsanity.com/paper/2310.01405/) (2023)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [BERT Rediscovers the Classical NLP Pipeline](https://mlsanity.com/paper/1905.05950/) (2019)
- [Mass-Editing Memory in a Transformer](https://mlsanity.com/paper/2210.07229/) (2022)
- [Attention is not Explanation](https://mlsanity.com/paper/1902.10186/) (2019)
- [Chameleon: Mixed-Modal Early-Fusion Foundation Models](https://mlsanity.com/paper/2405.09818/) (2024)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [Show-o: One Single Transformer to Unify Multimodal Understanding and Generation](https://mlsanity.com/paper/2408.12528/) (2024)
- [Towards Automated Circuit Discovery for Mechanistic Interpretability](https://mlsanity.com/paper/2304.14997/) (2023)
- [Network Dissection: Quantifying Interpretability of Deep Visual Representations](https://mlsanity.com/paper/1704.05796/) (2017)
- [Multimodal Few-Shot Learning with Frozen Language Models](https://mlsanity.com/paper/2106.13884/) (2021)

## Built on by (most cited first, 30 of 34)

- [Locate, Steer, and Improve: A Practical Survey of Actionable Mechanistic Interpretability in Large Language Models](https://mlsanity.com/paper/2601.14004/) (2026)
- [LLM-Powered AI Agent Systems and Their Applications in Industry](https://mlsanity.com/paper/2505.16120/) (2025)
- [DocPruner: A Storage-Efficient Framework for Multi-Vector Visual Document Retrieval via Adaptive Patch-Level Embedding Pruning](https://mlsanity.com/paper/2509.23883/) (2025)
- [BiasEdit: Debiasing Stereotyped Language Models via Model Editing](https://mlsanity.com/paper/2503.08588/) (2025)
- [Vision Function Layer in Multimodal LLMs](https://mlsanity.com/paper/2509.24791/) (2025)
- [Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation](https://mlsanity.com/paper/2509.22496/) (2025)
- [Mechanistic Finetuning of Vision-Language-Action Models via Few-Shot Demonstrations](https://mlsanity.com/paper/2511.22697/) (2025)
- [Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring](https://mlsanity.com/paper/2512.12069/) (2025)
- [V-SEAM: Visual Semantic Editing and Attention Modulating for Causal Interpretability of Vision-Language Models](https://mlsanity.com/paper/2509.14837/) (2025)
- [Fine-Grained Post-Training Quantization for Large Vision Language Models with Quantization-Aware Integrated Gradients](https://mlsanity.com/paper/2603.17809/) (2026)
- [GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs](https://mlsanity.com/paper/2507.18043/) (2025)
- [Evaluation of Geolocation Capabilities of Multimodal Large Language Models and Analysis of Associated Privacy Risks](https://mlsanity.com/paper/2506.23481/) (2025)
- [Do Language Models Associate Sound with Meaning? A Multimodal Study of Sound Symbolism](https://mlsanity.com/paper/2511.10045/) (2025)
- [Does AI See like Art Historians? Interpreting How Vision Language Models Recognize Artistic Style](https://mlsanity.com/paper/2603.11024/) (2026)
- [I Spy With My Model's Eye: Visual Search as a Behavioural Test for MLLMs](https://mlsanity.com/paper/2510.19678/) (2025)
- [SPD-Faith Bench: Diagnosing and Improving Faithfulness in Chain-of-Thought for Multimodal Large Language Models](https://mlsanity.com/paper/2602.07833/) (2026)
- [The physics of AI weather models](https://mlsanity.com/paper/2605.23778/) (2026)
- [MATEX: Multi-scale Attention and Text-guided Explainability of Medical Vision-Language Models](https://mlsanity.com/paper/2601.11666/) (2026)
- [Backdoor Directions in Vision Transformers](https://mlsanity.com/paper/2603.10806/) (2026)
- [From Heads to Neurons: Causal Attribution and Steering in Multi-Task Vision-Language Models](https://mlsanity.com/paper/2604.17941/) (2026)
- [Do Factual Recall Mechanisms Carry over from Text to Speech in Multimodal Language Models?](https://mlsanity.com/paper/2605.22170/) (2026)
- [Which Modality Decides? Counterfactual Modality Attribution for Multimodal LLMs](https://mlsanity.com/paper/2608.00076/) (2026)
- [Beyond the Linear Separability Ceiling: Aligning Representations in VLMs](https://mlsanity.com/paper/2507.07574/) (2025)
- [Interpret, prune and distill Donut : towards lightweight VLMs for VQA on document](https://mlsanity.com/paper/2509.26235/) (2025)
- [Understanding Temporal Logic Consistency in Video-Language Models through Cross-Modal Attention Discriminability](https://mlsanity.com/paper/2510.08138/) (2025)
- [Structural Graph Probing of Vision-Language Models](https://mlsanity.com/paper/2603.27070/) (2026)
- [Breaking Bad: Interpretability-Based Safety Audits of State-of-the-Art LLMs](https://mlsanity.com/paper/2604.20945/) (2026)
- [SimReg: Achieving Higher Performance in the Pretraining via Embedding Similarity Regularization](https://mlsanity.com/paper/2605.08809/) (2026)
- [Unsupervised Features Mining via Activation Geometry](https://mlsanity.com/paper/2607.04222/) (2026)
- [Through the LENS: Local Geometric Decomposition of Vision-Language Model Representations](https://mlsanity.com/paper/2608.00561/) (2026)

## Abstract (arXiv)

The rise of foundation models has transformed machine learning research, prompting efforts to uncover their inner workings and develop more efficient and reliable applications for better control. While significant progress has been made in interpreting Large Language Models (LLMs), multimodal foundation models (MMFMs) - such as contrastive vision-language models, generative vision-language models, and text-to-image models - pose unique interpretability challenges beyond unimodal frameworks. Despite initial studies, a substantial gap remains between the interpretability of LLMs and MMFMs. This survey explores two key aspects: (1) the adaptation of LLM interpretability methods to multimodal models and (2) understanding the mechanistic differences between unimodal language models and crossmodal systems. By systematically reviewing current MMFM analysis techniques, we propose a structured taxonomy of interpretability methods, compare insights across unimodal and multimodal architectures, and highlight critical research gaps.