arXiv:2602. 12279v2 Announce Type: replace-cross Abstract: Unified models can handle both multimodal understanding and generation within a single architecture, yet they typically operate in a single pass without iteratively refining their outputs.
Metric questions about video require vision-language models to use supplied real-world references to convert visual measurements into physical units. Yet we find that current models use this scale inf...
arXiv:2609.13158v1 Announce Type: new Abstract: Large Vision--Language Models (LVLMs) are increasingly expected to perform visual question answering (VQA) over planar media. However, existing planar...
arXiv:2610.19443v1 Announce Type: new Abstract: Native multimodal pre-training trains a single model from scratch on text and multimodal data, enabling it to understand both language and visual input...
The paper investigates how to scale a Transformer-based vision‑language model trained from scratch on multimodal inputs under a fixed compute budget. It finds that minimal loss follows a predictable compute law while optimal model size and token count obey power‑law scaling, with language and multimodal objectives showing distinct allocation trends. The study also proposes a budget‑compensation rule that allows additional multimodal tokens to offset text‑efficiency penalties, and demonstrates that native multimodal pre‑training improves spatial reasoning and few‑shot learning.
arXiv:2511. 21397v2 Announce Type: replace-cross Abstract: How does irrelevant information (i.
arXiv:2511. 18121v2 Announce Type: replace-cross Abstract: While Multimodal Large Language Models (MLLMs) excel on benchmarks, their processing paradigm differs from the human ability to integrate visual information.
The paper investigates why multimodal large language models (MLLMs) struggle with vision‑centric tasks when visual evidence conflicts with pretrained language knowledge. Using image reconstruction and a new WhatIfVis benchmark, the authors show that MLLMs preserve coarse‑grained visual attributes but fail to consistently use them, and that supervised fine‑tuning and activation patching can improve controllability of visual context sensitivity. The study demonstrates that the main bottleneck lies in the models’ inability to reliably regulate their reliance on visual evidence rather than in visual perception itself.
The paper introduces the Separable Law, a framework that predicts how vision‑language model performance varies with language backbone size and visual token count. By fitting this law to 26 InternVL and QwenVL models across a range of backbone sizes (1B–72B) and image resolutions (224–8K pixels), the authors show that some question types scale predictably with model capacity while others do not. The law also provides a closed‑form rule for allocating compute between backbone size and visual tokens, helping to choose near‑optimal model and image sizes under a fixed budget.
The paper introduces EquiSD, a label‑free training method that exploits scale equivariance to improve metric grounding in vision‑language models. By projecting model predictions onto a scale‑equivariant family and fine‑tuning on the resulting targets, EquiSD boosts a 3B model’s median response slope from 0.66 to 0.94 and raises mean relative accuracy by 9.2 points across simulated scales, with positive transfer to real QuantiPhy videos.
The paper introduces Percept-V, a dataset of 6,000 program-generated images across 30 domains designed to test simple visual perception skills from the TVPS-4 framework. Experiments show that state‑of‑the‑art multimodal large language models perform poorly compared to humans, especially as image complexity increases, and that fine‑tuning yields only limited generalization to related datasets. The study highlights specific perception skills that remain challenging for current models.
VKnowU is a benchmark that tests multimodal large language models (MLLMs) on their grasp of visual knowledge—intuitive, human-like understanding of physical and social principles in videos. The benchmark contains 1,680 questions across 1,249 videos, covering eight core types of visual knowledge, and shows that current state‑of‑the‑art MLLMs still lag behind human performance, especially on world‑centric tasks. To address this gap, the authors release VKnowQA and VideoKnow+, a baseline model that incorporates visual knowledge via a See‑Think‑Answer framework and reinforcement learning, improving performance on VKnowU and related datasets.