Induced Numerical Instability: Hidden Costs in Multimodal Large Language Models

• URL: http://arxiv.org/abs/2603.04453v1
• Date: Fri, 27 Feb 2026 18:47:36 GMT
• Title: Induced Numerical Instability: Hidden Costs in Multimodal Large Language Models
• Authors: Wai Tuck Wong, Jun Sun, Arunesh Sinha,
• Abstract summary: We study a novel mode of failure that causes degradation in performance indirectly by optimizing a loss term.<n>Our results uncover a fundamentally different vector of performance degradation, highlighting a failure mode not captured by adversarial perturbations.
• Score: 16.09514183229709
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: The use of multimodal large language models has become widespread, and as such the study of these models and their failure points has become of utmost importance. We study a novel mode of failure that causes degradation in performance indirectly by optimizing a loss term that seeks to maximize numerical instability in the inference stage of these models. We apply this loss term as the optimization target to construct images that, when used on multimodal large language models, cause significant degradation in the output. We validate our hypothesis on state of the art models large vision language models (LLaVa-v1.5-7B, Idefics3-8B, SmolVLM-2B-Instruct) against standard datasets (Flickr30k, MMVet, TextVQA, VQAv2, POPE, COCO) and show that performance degrades significantly, even with a very small change to the input image, compared to baselines. Our results uncover a fundamentally different vector of performance degradation, highlighting a failure mode not captured by adversarial perturbations.

Related papers

• Same Answer, Different Representations: Hidden instability in VLMs [65.36933543377346]
  We introduce a representation-aware and frequency-aware evaluation framework that measures internal embedding drift, spectral sensitivity, and structural smoothness.<n>We apply this framework to modern Vision Language Models (VLMs) across the SEEDBench, MMMU, and POPE datasets.
  arXiv  Detail & Related papers  (2026-02-06T12:24:26Z)
• Model-Dowser: Data-Free Importance Probing to Mitigate Catastrophic Forgetting in Multimodal Large Language Models [2.83595986479415]
  Fine-tuning Multimodal Large Language Models (MLLMs) on task-specific data is an effective way to improve performance on downstream applications.<n>Existing methods that aim to mitigate this issue either become ineffective when fine-tuning deeper layers of the language decoder or scale poorly with increasing model size.<n>We propose Model-Dowser, a novel sparse fine-tuning approach for MLLMs.
  arXiv  Detail & Related papers  (2026-02-04T12:56:27Z)
• Analyzing Diffusion and Autoregressive Vision Language Models in Multimodal Embedding Space [52.34072027212278]
  Embedding models are a fundamental component of modern AI systems such as semantic search and retrieval-augmented generation.<n>Recent advances in large foundation models have substantially accelerated the development of embedding models.<n>We present the first systematic study of converting Multimodal dLLMs into embedding models.
  arXiv  Detail & Related papers  (2026-01-19T06:51:15Z)
• LP-LLM: End-to-End Real-World Degraded License Plate Text Recognition via Large Multimodal Models [4.497411606350301]
  Real-world License Plate Recognition (LPR) faces significant challenges from severe degradations such as motion blur, low resolution, and complex illumination.<n>The prevailing "restoration-then-recognition" two-stage paradigm suffers from a fundamental flaw: the pixel-level optimization objectives of image restoration models are misaligned with the semantic goals of character recognition.<n>We propose an end-to-end structure-aware multimodal reasoning framework based on Qwen3-VL.
  arXiv  Detail & Related papers  (2026-01-14T03:32:55Z)
• Evaluating Robustness of Vision-Language Models Under Noisy Conditions [0.0176290054713643]
  Vision-Language Models (VLMs) have attained exceptional success across multimodal tasks such as image captioning and visual question answering.<n>We present a comprehensive evaluation framework to evaluate the performance of several state-of-the-art VLMs under controlled perturbations.
  arXiv  Detail & Related papers  (2025-09-15T22:31:21Z)
• Predictor-Corrector Enhanced Transformers with Exponential Moving Average Coefficient Learning [73.73967342609603]
  We introduce a predictor-corrector learning framework to minimize truncation errors. We also propose an exponential moving average-based coefficient learning method to strengthen our higher-order predictor. Our model surpasses a robust 3.8B DeepNet by an average of 2.9 SacreBLEU, using only 1/3 parameters.
  arXiv  Detail & Related papers  (2024-11-05T12:26:25Z)
• DEEM: Diffusion Models Serve as the Eyes of Large Language Models for Image Perception [66.88792390480343]
  We propose DEEM, a simple but effective approach that utilizes the generative feedback of diffusion models to align the semantic distributions of the image encoder.<n>DEEM exhibits enhanced robustness and a superior capacity to alleviate model hallucinations while utilizing fewer trainable parameters, less pre-training data, and a smaller base model size.
  arXiv  Detail & Related papers  (2024-05-24T05:46:04Z)
• Partially Recentralization Softmax Loss for Vision-Language Models Robustness [8.78222772167501]
  We study the adversarial robustness provided by modifying loss function of pre-trained multimodal models. Our experiments show that after a fine-tuning, adversarial robustness of pre-trained models can be significantly improved, against popular attacks.
  arXiv  Detail & Related papers  (2024-02-06T01:44:38Z)
• Identifying and Mitigating Model Failures through Few-shot CLIP-aided Diffusion Generation [65.268245109828]
  We propose an end-to-end framework to generate text descriptions of failure modes associated with spurious correlations. These descriptions can be used to generate synthetic data using generative models, such as diffusion models. Our experiments have shown remarkable textbfimprovements in accuracy ($sim textbf21%$) on hard sub-populations.
  arXiv  Detail & Related papers  (2023-12-09T04:43:49Z)
• Discrete Auto-regressive Variational Attention Models for Text Modeling [53.38382932162732]
  Variational autoencoders (VAEs) have been widely applied for text modeling. They are troubled by two challenges: information underrepresentation and posterior collapse. We propose Discrete Auto-regressive Variational Attention Model (DAVAM) to address the challenges.
  arXiv  Detail & Related papers  (2021-06-16T06:36:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.