PulseAugur
EN
LIVE 21:47:24

New research explores vision-language model advancements and safety

Multiple research papers published on arXiv explore advancements in vision-language models (VLMs). ViLFlow introduces a parallel generation approach for improved VQA and image captioning. MoMoVL focuses on modality-aware adaptive compression for on-device VLMs, particularly those using mixture-of-experts architectures. RePlan-VLA offers a framework for enhancing VLA models in long-horizon robotic manipulation tasks by replanning based on progress. Other research investigates the grounding capabilities of VLMs, the impact of preference optimization on object coverage, and scaling laws for vision-language fusion, alongside methods to diagnose and mitigate safety failures in these models. AI

IMPACT These papers advance VLM capabilities in areas like parallel generation, on-device efficiency, robotic control, and safety, potentially leading to more robust and versatile AI systems.

RANK_REASON Multiple research papers published on arXiv detailing new models and methodologies in vision-language processing.

Read on Hugging Face Daily Papers →

AI-generated summary · Google Gemini · from 46 sources. How we write summaries →

New research explores vision-language model advancements and safety

How we ranked this

Signal score
0 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
Multiple research papers published on arXiv detailing new models and methodologies in vision-language processing.
Source corroboration
46 independent sources
Strong cross-source corroboration — multiple independent publishers covered this within the clustering window.
Topics
paper, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
7 days old
Aged out of breaking-news scoring windows; ranking reflects the durable signal from the full source set.
Coverage growth since scoring
+15 source(s) since last score
New sources have picked up this story since our last re-score. Score will update on the next scoring pass.

Full methodology in our editorial standards.

COVERAGE [46]

  1. Hugging Face Daily Papers TIER_1 English(EN) ·

    TacInno: Teaching Vision-Language-Action Models the Contact Dynamics Beyond Vision and Action

    Vision-Language-Action (VLA) models have recently been equipped with tactile sensing to perceive visually subtle contact events in contact-rich manipulation. However, adding tactile input yields inconsistent gains, and some visuo-tactile VLA models rely little on tactile input, b…

  2. Hugging Face Daily Papers TIER_1 English(EN) ·

    Efficient Latent Reasoning for Vision-Language Model via Dynamics-Aware Quantization and Distillation

    Efficient chain-of-thought (CoT) reasoning in vision-language models (VLMs) requires reducing both intermediate reasoning steps and their execution cost. Latent CoT methods compress textual rationales into continuous thoughts, yet their full-precision (FP) models retain substanti…

  3. arXiv cs.AI TIER_1 English(EN) · Siddharth Vohra, Min Xu ·

    Know When to Zoom: How Small Is Too Small for Vision-Language Models?

    arXiv:2611.22709v1 Announce Type: cross Abstract: Online vision-language models often miss small details in large images, because the service shrinks the image, or rejects it, before the model ever sees it. A car 20 pixels wide in an 8000 by 6000 drone photo reaches Claude about …

  4. arXiv cs.AI TIER_1 English(EN) · Tongxi Wu, Yuyang Cheng, Jiacheng Xian, Menglin Yang, Kai Chen, Jian Zhang, Yang Gao ·

    Reading Scene States: Uncovering Safety Failures in Vision-Language Models

    arXiv:2611.23206v1 Announce Type: cross Abstract: Vision-language models (VLMs) can provide harmful assistance by interpreting visual events, even without an explicit request for the underlying harmful objective. We identify this \textbf{scene-state safety vulnerability} and intr…

  5. arXiv cs.AI TIER_1 English(EN) · Hansang Lee ·

    Say Less, Hallucinate Less? Diagnosing and Restoring Object Coverage in Preference-Optimized Vision-Language Models

    arXiv:2611.23179v1 Announce Type: cross Abstract: Preference optimization (PO) is the standard way to reduce object hallucination in large vision-language models. We find that four public PO checkpoints of LLaVA-1.5-7B also lower object coverage. They name fewer object categories…

  6. arXiv cs.LG TIER_1 English(EN) · Kevin Zhai, Bian Sun, Zhenyi Wang, Song Wang ·

    ViLFlow: Parallel Vision-Language Generation with Continuous Latent Flows

    arXiv:2611.24240v1 Announce Type: cross Abstract: Vision-language generation has predominantly relied on autoregressive and discrete diffusion models. Autoregressive models generate tokens sequentially, while discrete diffusion models typically predict tokens independently within…

  7. arXiv cs.LG TIER_1 English(EN) · Apoorva Vashisth, Manav Kulshrestha, Damon Conover, Zachary Kingston, Aniket Bera ·

    RePlan-VLA: Progress-Aware Replanning for Long-Horizon Vision-Language-Action Models

    arXiv:2611.23843v1 Announce Type: cross Abstract: Vision-Language-Action (VLA) models have emerged as a promising paradigm for general-purpose robotic manipulation, yet their performance often degrades on long-horizon tasks that require multiple interdependent manipulations. Inst…

  8. arXiv cs.LG TIER_1 English(EN) · Dongkeun Lee, Wonjun Lee ·

    MoMoVL: Modality-Aware Adaptive Compression for On-Device Sparsely-Activated Vision-Language Models

    arXiv:2611.23695v1 Announce Type: cross Abstract: The rise of the agentic web has been accompanied by growing demand for on-device Vision-Language Models (VLMs) that perform diverse cross-modal web tasks with enhanced privacy and reduced latency. Mixture-of-Experts (MoE) models a…

  9. arXiv cs.LG TIER_1 English(EN) · Michael Schroeder ·

    Where Global Context and Weight Sharing Work in Compact Vision Backbones

    arXiv:2611.23446v1 Announce Type: cross Abstract: Compact convolutional backbones such as EfficientNet achieve strong parameter efficiency through local depthwise separable convolutions, yet two structural questions remain central to hybrid design: where does global self-attentio…

  10. arXiv cs.LG TIER_1 English(EN) · Marcel Aach, Jiangtao Wang, Stefan Kesselheim, Andreas Lintermann ·

    OpenGNS: An Open Dataset of the Gradient Noise Scale across Vision, Language, and Diffusion Models

    arXiv:2611.24331v1 Announce Type: new Abstract: The Gradient Noise Scale (GNS) metric measures the signal-to-noise ratio of gradients in stochastic gradient descent and, as shown by McCandlish et al. (2018), can be used to approximate the Critical Batch Size (CBS) during neural n…

  11. arXiv cs.CL TIER_1 English(EN) · Basavaraj R M, Harshal Chandel ·

    The Illusion of Grounding: Do Vision-Language Models Weigh Retrieved Text and Images Like an Ideal Observer?

    arXiv:2611.23127v1 Announce Type: cross Abstract: Retrieval-augmented vision-language models (VLMs) answer from two sources of evidence that can each fail: the retrieved text can be wrong or noisy, and the image can be degraded or contradicted. We introduce GIB-200K, 200,000 imag…

  12. Hugging Face Daily Papers TIER_1 English(EN) ·

    Teaching Vision-Language-Action Models to Stop Safely

    Vision-language-action (VLA) models enable robots to perform diverse tasks from visual observations and language instructions. While prior work has explored task suppression and refusal, stopping prohibited behavior once execution is underway and controlling the resulting physica…

  13. Hugging Face Daily Papers TIER_1 English(EN) ·

    Probing the "Conscious" Mind: Selective Access and Global Workspace in Vision-Language Models

    Vision-language models (VLMs) integrate visual and textual information, but their outputs do not reveal internal concept representations. Global workspace theory (GWT) proposes that information is selected from parallel brain processes, integrated in a shared workspace, and made …

  14. Hugging Face Daily Papers TIER_1 English(EN) ·

    $Z$-sight: Unifying Vision-Language-Action and Latent World Modeling

    We introduce $Z$-sight, a unified robot foundation model that couples vision-language-action (VLA) and generative latent world modeling. Video-based world action models (WAMs) show the value of joint future-and-action generation for control, yet inherit the heavy cost of video ge…

  15. arXiv cs.LG TIER_1 English(EN) · Seng Pei Liew, Teppei Suzuki, Seiya Tokui ·

    Scaling Laws for Vision-Language Fusion

    arXiv:2611.22351v1 Announce Type: new Abstract: Vision-language models (VLMs) have emerged as central foundation models combining multimodal perception with text comprehension. While VLMs are traditionally built via late fusion by continued pre-training an off-the-shelf language …

  16. Hugging Face Daily Papers TIER_1 English(EN) ·

    RoughCut-Bench: A Benchmark for Automatic Video Rough-Cut with Large Vision-Language Models

    A new benchmark and agentic workflows assess vision-language models for automatic video rough-cut, revealing current models struggle to match expert cut plans.

  17. Hugging Face Daily Papers TIER_1 English(EN) ·

    Know When to Zoom: How Small Is Too Small for Vision-Language Models?

    Online vision-language models often miss small details in large images, because the service shrinks the image, or rejects it, before the model ever sees it. A car 20 pixels wide in an 8000 by 6000 drone photo reaches Claude about 6 pixels wide, and OpenAI's default rejects the ph…

  18. Hugging Face Daily Papers TIER_1 English(EN) ·

    Lost in Read-out: Locating the Image Quality Gap in Vision-Language Models with a Closed-Form Edit

    Vision-language models (VLMs) are widely used as zero-shot image quality assessors, yet their good-vs-poor answers agree only moderately with human opinion. We ask whether these models fail to see quality or fail to say it. Layer-wise probes on five VLMs point to the second: qual…

  19. Hugging Face Daily Papers TIER_1 English(EN) ·

    Perception into Imagination: Executable Memory for World Models

    Physical agents must use what they previously saw or touched after visibility or contact is lost. Their actions change the implications of these remembered facts. We introduce executable perceptual memory (EPM), a world-model framework that advances explicit perceptual state alon…

  20. Hugging Face Daily Papers TIER_1 English(EN) ·

    VSpark: Speculative Decoding for Autoregressive Vision-Language-Action Models

    VSpark accelerates autoregressive vision-language-action models by improving speculative decoding with calibrated action-noise rollouts, hidden-feature alignment, and reuse of rejected-draft continuations, achieving lossless 1.28x–2.10x speedups across multiple robot manipulation…

  21. Hugging Face Daily Papers TIER_1 English(EN) ·

    AnachroBench: Evaluating Historical Consistency in Large Vision-Language Models

    Existing Large Vision-Language Model benchmarks primarily evaluate visual understanding in general everyday scenarios, leaving historical consistency with a specified era and country largely underexplored. To address this gap, we propose AnachroBench, a controlled discriminative …

  22. Hugging Face Daily Papers TIER_1 English(EN) ·

    FAVR: Flow-Aware Visual Residual Alignment for Vision-Language-Action Policies

    Pretrained video representations provide rich temporal information for robotic manipulation, yet effectively integrating them into existing vision-language-action (VLA) policies remains challenging. We present Flow-Aware Visual Residual Alignment (FAVR), a lightweight framework f…

  23. arXiv cs.AI TIER_1 English(EN) · Zhihao Zhan, Jiaying Zhou, Likui Zhang, Qinhan Lyu, Hao Liu, Weizheng Li, Jusheng Zhang, Ziliang Chen, Tianshui Chen, Ruifeng Zhai, Keze Wang, Liang Lin, Guangrun Wang ·

    E0: Expressive Fine-Grained Discrete Action Prediction for Vision-Language-Action Models via Tweedie Discrete Diffusion

    arXiv:2511.21542v3 Announce Type: replace-cross Abstract: Vision-Language-Action (VLA) models offer a unified framework for robotic manipulation by integrating visual perception, language understanding, and control generation. However, existing VLA systems still struggle to gener…

  24. arXiv cs.AI TIER_1 English(EN) · Yucheng Guo, Yongjian Guo, Zhong Guan, Wen Huang, Haodong Yue, Shuai Di, Junwu Xiong, Yicheng Gong, Xi Xiao ·

    D-VLA: A High-Concurrency Distributed Asynchronous Reinforcement Learning Framework for Vision-Language-Action Models

    arXiv:2605.13276v3 Announce Type: replace Abstract: The rapid evolution of Embodied AI has enabled Vision-Language-Action (VLA) models to excel in multimodal perception and task execution. However, applying Reinforcement Learning (RL) to these massive models in large-scale distri…

  25. arXiv cs.AI TIER_1 English(EN) · Vaidehi Bagaria, Bijo Sebastian, Nirav Kumar Patel ·

    Recursive Belief Vision Language Action Models

    arXiv:2602.20659v3 Announce Type: replace Abstract: Vision-language-action models must enable agents to execute long-horizon tasks under partial observability. However, most existing approaches remain observation-driven, relying on short context windows or repeated queries to vis…

  26. arXiv cs.AI TIER_1 English(EN) · Suvarna Kadam (Witsense AI), Soran Parsa (University of Huddersfield) ·

    Vision-Language-Action Models: A Life-Cycle Survey of Design, Training and Deployment

    arXiv:2611.18372v1 Announce Type: cross Abstract: Vision-language-action (VLA) models adapt web-pretrained vision-language models to output robot actions, so that web knowledge carries over to physical control. In under 4 years the field has moved from RT-2, the first VLA, to ope…

  27. arXiv cs.AI TIER_1 English(EN) · Tianjun Gu, Jingyu Gong, Zhizhong Zhang, Yuan Xie, Lizhuang Ma, Xin Tan, Athanasios V. Vasilakos ·

    Vision-language models lag human performance on physical dynamics and intent reasoning

    arXiv:2601.01547v3 Announce Type: replace-cross Abstract: Spatial intelligence is central to embodied cognition, yet contemporary AI systems still struggle to reason about physical interactions in open-world human environments. Despite strong performance on controlled benchmarks,…

  28. Hugging Face Daily Papers TIER_1 English(EN) ·

    ScaleVLA: Scaling Vision-Language-Action Models to 552 Billion Parameters

    Frontier Vision-Language Models (VLMs) such as GPT-6 Astra have demonstrated remarkable capabilities in robot control, suggesting that broad pretraining provides knowledge valuable for physical action. Yet most Vision-Language-Action (VLA) policies still rely on much smaller VLM …

  29. arXiv cs.AI TIER_1 English(EN) · Wen-Han Hsieh, Ken Jiankun Zheng, Chia-Chien Chen, David M. Chan, Trevor Darrell ·

    Learning Low-Latency Action Verification for Frozen Vision-Language-Action Models

    arXiv:2611.14834v1 Announce Type: cross Abstract: Vision-language-action (VLA) models have made significant progress in general-purpose robotic manipulation. However, sampled action sequences can be ineffective, while evaluating alternatives with computationally expensive verifie…

  30. arXiv cs.CV TIER_1 English(EN) · Zhiming Luo, Hebaixu Wang, Haonan Guo, Jing Zhang, Bo Du ·

    VLRS-Bench: A Vision-Language Reasoning Benchmark for Remote Sensing

    arXiv:2602.07045v3 Announce Type: replace Abstract: Recent advancements in Multimodal Large Language Models (MLLMs) have enabled complex reasoning. However, existing remote sensing (RS) benchmarks remain heavily biased toward perception tasks, such as object recognition and scene…

  31. arXiv cs.CV TIER_1 English(EN) · Xue Wang, Yubin Xu, Kun Tan ·

    RICapBench: A Fine-Grained Visual Understanding Benchmark Dataset for Vision-Language Models in Remote Sensing

    arXiv:2611.16413v2 Announce Type: replace Abstract: Advances in vision-language models (VLMs) have improved remote sensing image analysis and supported geospatial decision-making. However, fine-grained understanding of remote sensing imagery remains challenging for VLMs. Large va…

  32. arXiv cs.CV TIER_1 English(EN) · Dingqiang Ye, Jiacong Xu, Wen Zhang, Yiqun Mei, Vishal M. Patel ·

    Recurrent Vision Language Models Solve the Shell Game

    arXiv:2611.25210v1 Announce Type: new Abstract: Tracking a hidden object through a video is effortless for humans, yet it remains a bottleneck for vision language models (VLMs). In the shell game, where a ball is hidden under one of several identical cups that are then swapped, V…

  33. arXiv cs.CV TIER_1 English(EN) · Jiajun Fan, Neil He, Bangji Yang, Qinglong Zheng, Ziwen Wang, Chaoran Cheng, Changyi Yang, Hanqi Mao, Ge Liu ·

    Keel: A Frontier Model Teaches a Vision-Language Model by Rewriting Its Weights

    arXiv:2611.25257v1 Announce Type: new Abstract: Can a frontier model repair a vision-language model (VLM) by rewriting its weights? Small reasoning VLMs such as Qwen3.5 need such a repair: in thinking mode they fall into loops on many image questions, repeating themselves until t…

  34. arXiv cs.CV TIER_1 English(EN) · Mingda Zhang, Yuanhao Dong, Qiang Huang, Rui Qian, Yanjin Li, Zhenyu Liu, Xiao Huang, Zihao Yang, Jian Wang, Pinjia He, Xiaoying Tang ·

    BaseRSI: Recursive Self-Improvement of a Base Vision-Language Model via Training-Free Harness Internalization

    arXiv:2611.24578v1 Announce Type: new Abstract: Recursive self-improvement (RSI) is moving from the harness of prompts and tools around a base model toward the model itself. However, existing RSI still faces key challenges, including harness-bound gains that stay outside the base…

  35. arXiv cs.CV TIER_1 English(EN) · Aditya Gandhamal, Aniruddh Sikdar, Arya Kishor, Suresh Sundaram ·

    Pixel2Perspective: A Benchmark for Multi-Modal and Multi-View Generalization in Vision-Language Models

    arXiv:2611.24327v1 Announce Type: new Abstract: Embodied AI systems, such as drones and autonomous vehicles, improve navigation by enhancing perception in environments containing unseen objects. While Visual Language Models (VLMs) offer strong image understanding, their real-worl…

  36. arXiv cs.CV TIER_1 English(EN) · Zhenyu Li, Tianyi Shang, Shuhan Dong, Ruotong Xu, Qingyu Zhu ·

    AcrossPlace: Do Vision--Language Models Recognize Places Beyond Their Source Views? A Comprehensive Benchmark

    arXiv:2611.15314v2 Announce Type: replace Abstract: Language-guided place recognition requires recovering geographic identity across visual observations beyond the image from which a description originates. However, aggregate retrieval scores often entangle observation transfer, …

  37. arXiv cs.CV TIER_1 English(EN) · Junjie Ye, Yiqi Zhao, Muchen Xu, Celina Shiyu Wang, Hao Dong, Ruohai Ge, Rong Xue, Benjamin J Kolligs, Aditya Dutt, Sasanka Nagavalli, Nassim A. Samad, Gaurav S. Sukhatme, Jyotirmoy V. Deshmukh, Yue Wang, Yuliang Guo ·

    $Z$-sight: Unifying Vision-Language-Action and Latent World Modeling

    arXiv:2611.24940v1 Announce Type: cross Abstract: We introduce $Z$-sight, a unified robot foundation model that couples vision-language-action (VLA) and generative latent world modeling. Video-based world action models (WAMs) show the value of joint future-and-action generation f…

  38. arXiv cs.CV TIER_1 English(EN) · Wenhao Li, Zhongyi Han, Yuhao Wei, Liang Han, Yilong Yin, Liqiang Nie ·

    Towards Pose-Robust Few-Shot Vision-Language-Action Models

    arXiv:2611.24360v1 Announce Type: cross Abstract: Fine-tuning a vision-language-action (VLA) model on a few dozen demonstrations is the standard way to put a manipulation policy on a robot, yet the resulting policy silently assumes that the camera and the robot base stay exactly …

  39. arXiv cs.CV TIER_1 English(EN) · Xinyi Guo, Kaile Huang, Xitong Ling, Mingyi He, Chende Zheng, Jianyu Xu, Chao He, Dongqing Song ·

    Read Together, Answer Apart: How Earlier Questions Help Vision-Language Models Look

    arXiv:2611.24785v1 Announce Type: new Abstract: Vision-language models (VLMs) are often asked several questions about the same image in a single prompt (co-asking), yet existing evaluations report conflicting accuracy gains and hallucinations at the prompt level without tracking …

  40. arXiv cs.CV TIER_1 English(EN) · Ruisi Zhao, You Li, Sihan Chen, Aosong Cheng, Chixiang Ma, Sen Pei, Zhen Yang, Hehe Fan, Dongliang He, Yi Yang ·

    RoughCut-Bench: A Benchmark for Automatic Video Rough-Cut with Large Vision-Language Models

    arXiv:2611.24098v1 Announce Type: new Abstract: Video rough-cut is a core stage in media production workflows, organizing raw media materials into an initial video cut. Despite recent advances in large vision-language models (LVLMs) for video-related tasks, automatic rough-cut re…

  41. arXiv cs.CV TIER_1 English(EN) · Avinash Goutham Aluguvelly ·

    Weight Space Programming for Vision Language Model Repair

    arXiv:2611.23615v1 Announce Type: new Abstract: In weight-space programming, a language model edits a network's weights directly, without gradients. We ask what a successful repair of this kind relies on. We plant faults in the weights of two vision-language models, SmolVLM-256M …

  42. arXiv cs.CV TIER_1 English(EN) · Hui Zhu, Saleh Zare Zade, Rafi Ibn Sultan, Xiangyu Zhou, Dongxiao Zhu ·

    Beyond the Sink Token: Sink-Guided Key Editing in Vision-Language Models

    arXiv:2611.23516v1 Announce Type: new Abstract: In vision-language models (VLMs), a few sink tokens attract a disproportionate share of attention. While previous studies have examined why these tokens attract attention, we investigate how the sink-key direction affects attention …

  43. arXiv cs.CV TIER_1 English(EN) · Ling Lin, Heng Su, Xiangzeng Liu, Yang Bai, Congcong Zhu, Jingrun Chen ·

    Can Vision-Language Models Truly Understand Images? A Benchmark Beyond Commonsense Bias

    arXiv:2611.22360v1 Announce Type: new Abstract: Multimodal large language models (MLLMs) have achieved impressive performance on a wide range of vision-language tasks. However, it remains unclear whether their predictions are grounded in the visual content itself or are largely d…

  44. arXiv cs.CV TIER_1 English(EN) · Qixuan Wu (School of Artificial Intelligence, Beihang University, Beijing, China), Xiaole Yang (School of Software, Beihang University, Beijing, China), Yuguang Yang (School of Electronic and Information Engineering, Beihang University, Beijing, China), … ·

    FAVR: Flow-Aware Visual Residual Alignment for Vision-Language-Action Policies

    arXiv:2611.20971v2 Announce Type: replace Abstract: Pretrained video representations provide rich temporal information for robotic manipulation, yet effectively integrating them into existing vision-language-action (VLA) policies remains challenging. We present Flow-Aware Visual …

  45. arXiv cs.CV TIER_1 English(EN) · Zhuangzi Li, Shilv Cai, Jian Jin, Zhiyi Pan, Haihong Xiao, Weisi Lin ·

    Lost in Read-out: Locating the Image Quality Gap in Vision-Language Models with a Closed-Form Edit

    arXiv:2611.22564v1 Announce Type: new Abstract: Vision-language models (VLMs) are widely used as zero-shot image quality assessors, yet their good-vs-poor answers agree only moderately with human opinion. We ask whether these models fail to see quality or fail to say it. Layer-wi…

  46. Mastodon — mastodon.social TIER_1 English(EN) · notatechguy ·

    Vision-language models selectively access internal concepts An arXiv preprint finds vision-language models selectively access concept information as tasks requi

    Vision-language models selectively access internal concepts An arXiv preprint finds vision-language models selectively access concept information as tasks require, echoing global workspace theory from https://www. notatechguy.com/vision-languag e-models-selectively-access-interna…