# Understanding Multimodal LLMs Under Distribution Shifts: An Information-Theoretic Approach

arXiv: https://arxiv.org/abs/2502.00577 · HTML: https://mlsanity.com/paper/2502.00577/

## Key facts

- Cited by 16 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 56 of its 80 references that are in the map.
- First posted on arXiv 1 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, 30 of 56)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [Training language models to follow instructions with human feedback](https://mlsanity.com/paper/2203.02155/) (2022)
- [Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena](https://mlsanity.com/paper/2306.05685/) (2023)
- [Visual Instruction Tuning](https://mlsanity.com/paper/2304.08485/) (2023)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuning](https://mlsanity.com/paper/2305.06500/) (2023)
- [MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Models](https://mlsanity.com/paper/2304.10592/) (2023)
- [CLIPScore: A Reference-free Evaluation Metric for Image Captioning](https://mlsanity.com/paper/2104.08718/) (2021)
- [Evaluating Object Hallucination in Large Vision-Language Models](https://mlsanity.com/paper/2305.10355/) (2023)
- [Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scaling](https://mlsanity.com/paper/2412.05271/) (2024)
- [LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day](https://mlsanity.com/paper/2306.00890/) (2023)
- [Deep Variational Information Bottleneck](https://mlsanity.com/paper/1612.00410/) (2016)
- [InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets](https://mlsanity.com/paper/1606.03657/) (2016)
- [Robust fine-tuning of zero-shot models](https://mlsanity.com/paper/2109.01903/) (2021)
- [Fine-Tuning can Distort Pretrained Features and Underperform Out-of-Distribution](https://mlsanity.com/paper/2202.10054/) (2022)
- [BARTScore: Evaluating Generated Text as Text Generation](https://mlsanity.com/paper/2106.11520/) (2021)
- [MINE: Mutual Information Neural Estimation](https://mlsanity.com/paper/1801.04062/) (2018)
- [Prometheus: Inducing Fine-grained Evaluation Capability in Language Models](https://mlsanity.com/paper/2310.08491/) (2023)
- [On Variational Bounds of Mutual Information](https://mlsanity.com/paper/1905.06922/) (2019)
- [Opening the Black Box of Deep Neural Networks via Information](https://mlsanity.com/paper/1703.00810/) (2017)
- [RewardBench: Evaluating Reward Models for Language Modeling](https://mlsanity.com/paper/2403.13787/) (2024)
- [Measuring Robustness to Natural Distribution Shifts in Image Classification](https://mlsanity.com/paper/2007.00644/) (2020)
- [On Mutual Information Maximization for Representation Learning](https://mlsanity.com/paper/1907.13625/) (2019)
- [CLUB: A Contrastive Log-ratio Upper Bound of Mutual Information](https://mlsanity.com/paper/2006.12013/) (2020)

## Built on by (most cited first, 16 of 16)

- [Uncertainty Quantification in LLM Agents: Foundations, Emerging Challenges, and Opportunities](https://mlsanity.com/paper/2602.05073/) (2026)
- [Understanding Language Prior of LVLMs by Contrasting Chain-of-Embedding](https://mlsanity.com/paper/2509.23050/) (2025)
- [SPINE: Token-Selective Test-Time Reinforcement Learning with Entropy-Band Regularization](https://mlsanity.com/paper/2511.17938/) (2025)
- [Failure Modes in LLM Systems: A System-Level Taxonomy for Reliable AI Applications](https://mlsanity.com/paper/2511.19933/) (2025)
- [UniGame: Turning a Unified Multimodal Model Into Its Own Adversary](https://mlsanity.com/paper/2511.19413/) (2025)
- [Visual Instruction Bottleneck Tuning](https://mlsanity.com/paper/2505.13946/) (2025)
- [A Comprehensive Information-Decomposition Analysis of Large Vision-Language Models](https://mlsanity.com/paper/2603.29676/) (2026)
- [Adapter-Augmented Bandits for Online Multi-Constrained Multi-Modal Inference Scheduling](https://mlsanity.com/paper/2603.06403/) (2026)
- [Are Video Reasoning Models Ready to Go Outside?](https://mlsanity.com/paper/2603.10652/) (2026)
- [ICYM2I: The illusion of multimodal informativeness under missingness](https://mlsanity.com/paper/2505.16953/) (2025)
- [Beyond In-Domain Detection: SpikeScore for Cross-Domain Hallucination Detection](https://mlsanity.com/paper/2601.19245/) (2026)
- [NeuroVLM-Bench: Evaluation of Vision-Enabled Large Language Models for Clinical Reasoning in Neurological Disorders](https://mlsanity.com/paper/2603.24846/) (2026)
- [Geometry-Aware Uncertainty Coresets for Robust Visual In-Context Learning in Histopathology](https://mlsanity.com/paper/2605.18419/) (2026)
- [CoDA: Towards Effective Cross-domain Knowledge Transfer via CoT-guided Domain Adaptation](https://mlsanity.com/paper/2604.19488/) (2026)
- [COMPASS: COntinual Multilingual PEFT with Adaptive Semantic Sampling](https://mlsanity.com/paper/2604.20720/) (2026)
- [Local Sparsity Enables Unsupervised LLM Safety Detection](https://mlsanity.com/paper/2609.20129/) (2026)

## Abstract (arXiv)

Multimodal large language models (MLLMs) have shown promising capabilities but struggle under distribution shifts, where evaluation data differ from instruction tuning distributions. Although previous works have provided empirical evaluations, we argue that establishing a formal framework that can characterize and quantify the risk of MLLMs is necessary to ensure the safe and reliable application of MLLMs in the real world. By taking an information-theoretic perspective, we propose the first theoretical framework that enables the quantification of the maximum risk of MLLMs under distribution shifts. Central to our framework is the introduction of Effective Mutual Information (EMI), a principled metric that quantifies the relevance between input queries and model responses. We derive an upper bound for the EMI difference between in-distribution (ID) and out-of-distribution (OOD) data, connecting it to visual and textual distributional discrepancies. Extensive experiments on real benchmark datasets, spanning 61 shift scenarios, empirically validate our theoretical insights.