Vision-Language Grounding

Visual grounding aims to locate image regions that correspond to natural language descriptions and is a key component of interpretable vision systems. In remote sensing imagery, grounding is particularly challenging due to complex scenes, small objects, and large variations in scale. Relying on a single model is often insufficient to address these diverse challenges. In this work, we propose two grounding pipelines, Sequential Grounding Refinement (SGR) and Cluster-Aware Grounding Refinement (CGR), that combine the complementary strengths of RemoteSAM, a visual grounding model specialized for remote sensing, and SAM3, a powerful general-purpose segmentation model. Our approach first uses RemoteSAM to obtain an initial estimate of object location, which is then refined using SAM3 to produce more accurate and spatially consistent segmentations. Additionally, we explore an ensemble strategy based on majority voting across six diverse grounding pipelines, each with distinct capabilities. This multi-model framework improves robustness and significantly enhances localization accuracy. Experimental results demonstrate that the proposed pipelines and ensemble approach outperform individual models, leading to more reliable and precise visual grounding predictions.

Robots operating in diverse environments rely on visual input to interpret objects and spatial layouts. In human-collaborative tasks, they are expected to communicate this understanding through language. Vision-language models (VLMs) support robotic tasks involving visual interpretation, question answering, and instruction following, but their capabilities in collaborative dialogue tasks requiring spatial reasoning remain underexplored. We study this gap through a collaborative structure-building task that combines visual interpretation, grounding, language-guided interaction, and action generation. We develop a framework in which VLMs use dialogue to reconstruct a target structure from visual and textual inputs. We evaluate open-weight and closed VLMs across interaction settings, input modalities, and image representations. Results show that spatial reasoning over visual representations remains difficult for the evaluated VLMs. Detailed text representations of the target yield higher reconstruction success across modality conditions, while decomposed image representations improve performance. These findings reveal limits in visual spatial grounding and grounded instruction generation for collaborative VLM agents.

In-context localization (ICL) seeks to localize a target object specified by a small set of support examples in a query image, operating on the fly without training or parameter updates. Despite rapid advances in vision-language models (VLMs), achieving category-agnostic and visually grounded ICL remains an open problem, even though it is essential for applications such as image editing, personalized visual search, and retrieval. Existing methods are fragile and rely on explicit category supervision, which not only limits applicability in realistic settings with unnamed or instance-specific objects but also introduces category bias that steers predictions toward semantic priors rather than visual evidence. We introduce a two-stage training framework that explicitly optimizes in-context attention between support bounding boxes and query images without category supervision. We further refine localization via reinforcement learning using Group Relative Policy Optimization (GRPO) to directly minimize localization error. This formulation enforces visual correspondence over semantic priors, yielding robust instance-level localization. Empirically, a 7B-parameter model trained with our objectives outperforms models up to 72B parameters, demonstrating that context-aware localization objectives can surpass scaling alone. Comprehensive ablations validate the contribution of each component.

While visually grounded Chain-of-Thought (CoT) has emerged as a promising paradigm to enhance fine-grained perception in multimodal large language models (MLLMs), its efficacy during the inference phase remains underexplored. In this work, we empirically find that mandating explicit object boxes in visually grounded CoT during inference often degrades performance compared to standard textual CoT, which reasons without explicit visual grounding. We hypothesize that the visual localization capability can be internalized into the textual CoT and that the mandatory explicit grounding introduces unnecessary interference with the model's primary objective of answer prediction. To address this problem, we propose Internalizing Visually Grounded Reasoning (\textbf{iVGR}), a novel reinforcement learning framework that transfers localization capabilities into the textual reasoning process. We employ a dual-stream training strategy, where a textual stream is aligned with a high-quality visually grounded stream via a proposed consistency reward, enabling the model to localize accurately without explicit grounding during inference. Extensive experiments demonstrate that our method significantly outperforms existing baselines on fine-grained benchmarks, while maintaining the flexibility to support tool-assisted inference workflows.

Chain-of-thought (CoT) reasoning has significantly improved the reasoning ability of large vision-language models (LVLMs) by verbalizing intermediate reasoning steps in natural language. However, such discrete textual rationales are often insufficient for encoding continuous visual evidence. Recent work addresses this limitation by moving reasoning into continuous latent space. Despite promising progress, existing methods leave latent reasoning insufficiently connected to the compositional and relational structure of visual evidence. To address this gap, we introduce ReGuLaR, a relation grounded latent reasoning framework that explicitly grounds latent states in these critical yet overlooked visual evidence. ReGuLaR uses a training-time ReGFormer to focus latent reasoning on question-relevant objects and inter-object relations, while at inference time the model reasons and generates answers without invoking the ReGFormer. To support training ReGuLaR, we construct RGROUNDING-351K, a real-world vision-language dataset annotated with key object bounding boxes and inter-object relations. Extensive experiments across diverse benchmarks show that ReGuLaR consistently outperforms existing approaches and achieves state-of-the-art performance. We include our code in the submission and will release the code and training data publicly upon acceptance.

Large vision-language models (LVLMs) often hallucinate objects that are not present in the input image, largely because visual grounding weakens as decoding progresses. Existing inference-time mitigation methods modify logits or hidden states throughout generation, but they suffer from three key limitations: they lack an explicit grounding objective, intervene even when the model is already well-grounded, and use fixed correction strengths that do not adapt to the severity of grounding failure. We propose BRACS (Barrier-Regulated Adaptive Closed-form Steering), a training-free steering framework that addresses these issues through barrier-regulated adaptive closed-form steering. BRACS monitors the model's own attention to measure visual grounding and applies corrections to the hidden states only when grounding deteriorates. The corrective update is computed analytically in closed form, requiring no training of auxiliary networks or model retraining. Experiments on LLaVA-1.5-7B and Qwen-VL-Chat show that BRACS consistently outperforms prior methods on hallucination benchmarks, reducing CHAIR$_s$ by 9.4 points and improving POPE F1 by 2.7 points, while matching or improving performance on four general multimodal benchmarks. BRACS also remains efficient, operating at 80% of greedy decoding throughput and achieving 1.3 times higher speed on average than the baselines.

Understanding long-form egocentric videos remains challenging for multimodal large language models (MLLMs) due to limited context length and insufficient grounding of fine-grained visual details. The recently proposed HD-EPIC benchmark highlights these limitations: even strong long-context models achieve relatively low performance across diverse video question answering tasks. In this paper, we propose a unified framework that decouples long-video reasoning into two complementary forms of evidence: semantic evidence and visual evidence. Semantic evidence captures global procedural structure through a coarse-to-fine extraction pipeline, while object-centric visual evidence preserves fine-grained grounding through bounding boxes and visual embeddings. During inference, we formulate reasoning as a query-conditioned evidence retrieval and integration process, dynamically selecting relevant information from both sources. Our approach achieves competitive performance in the HD-EPIC-VQA Challenge across multiple task categories. More broadly, our results demonstrate that explicitly structuring, retrieving, and integrating semantic and visual evidence is critical for effective long-video understanding with MLLMs.

Recent work has shown that Vision-Language Models (VLMs) used for optical character recognition (OCR) can generate plausible but visually unsupported text, suggesting reliance on language priors. Comparing open-weight VLMs with traditional OCR baselines on low-resource Ancient Greek critical editions, we show that VLM errors often remain fluent even when wrong, producing plausible Greek substitutions where traditional engines produce local recognition noise. To analyze visual evidence during decoding, we introduce controlled image perturbations and token-level grounding measures based on conditional versus image-free decoding distributions. Under character-level perturbations, VLMs diverge sharply from the perturbed ground truth while traditional OCR remains comparatively faithful; however, token-level analysis shows that prior reliance is model-specific: in an OCR-specialist model, fluent lexical errors are produced with little reliance on the image, whereas general-purpose VLMs remain conditioned on the visual input even when wrong. Decode-time interventions fail to reliably restore grounding, while post-OCR language-model correction improves several systems only by repairing text after generation. Our results extend prior evidence of OCR language-prior reliance to low-resource historical documents and a broader set of models, showing that fluent output is not necessarily visually grounded and motivating interpretability-driven evaluation beyond aggregate accuracy.

Vision-language models (VLMs) commonly formulate visual grounding and detection as a coordinate-token generation problem, serializing each 2D box into multiple 1D tokens that are learned and decoded largely independently. This token-by-token decoding mismatches the coupled structure of box geometry and creates a practical inference bottleneck due to strictly sequential generation. We introduce LocateAnything, a unified generative grounding and detection framework based on Parallel Box Decoding (PBD). By decoding geometric elements such as bounding boxes and points as atomic units in a single step, LocateAnything preserves intra-box geometric coherence and unlocks substantial parallelism. We show that PBD improves both decoding throughput and localization accuracy. We further develop a scalable data engine and curate LocateAnything-Data, a large-scale dataset with more than 138 million training samples, substantially increasing data diversity for high-precision localization. Extensive evaluations show that LocateAnything advances the speed-accuracy frontier, achieving significantly higher decoding throughput while improving high-IoU localization quality across diverse benchmarks. The results highlight the complementary benefits of Parallel Box Decoding and large-scale training data in enabling efficient and precise unified visual grounding and detection.

Modern referring image segmentation pipelines couple a vision-language model (VLM) for grounding with a promptable segmenter such as the Segment Anything Model (SAM) for mask generation. Prior training-free instances of this recipe consistently trail fine-tuned and reinforcement-learning (RL)-tuned specialists, and it has been unclear whether the gap comes from the VLM's grounding, SAM's capacity, or the prompt. We show that the gap is dominated by prompt ambiguity: a VLM-proposed bounding box (bbox) leaves SAM to guess which pixels inside the bbox belong to the object the expression denotes. Interior points are the natural disambiguator, but where they fall matters; prior work relies on naively sampled points that land on boundaries, distractors, and background clutter, and can even hurt performance compared to the bbox alone. Supervised and RL-tuned methods close this gap by training a VLM to predict better points; we show that this training is unnecessary. At a matched budget of five interior points, replacing naive sampling with stable, informative point selection improves cumulative Intersection-over-Union (cIoU) by 12-18 points across RefCOCO/+/g, with every model fixed. We turn this observation into PinPoint, a deterministic, training-free point selector that fuses four visual cues into a consensus map, selects compact, spatially diverse points away from boundaries, and uses the frozen VLM to label each point. Without any task-specific training, PinPoint matches supervised and RL-tuned specialists on the same stack while issuing only two VLM calls per query.

Understanding human gaze behavior is essential for complex scene comprehension and human-computer interaction. Traditional gaze following models are typically restricted to pure spatial localization, lacking the high-level capacity to reason about semantic targets or complex social contexts. Furthermore, these models often process individuals sequentially, requiring redundant computations over the same scene image for multi-person inference. While recent Vision-Language Models (VLMs) offer the exceptional semantic reasoning needed to address gaze-related semantic tasks, their reliance on discrete text generation inherently limits precision in continuous spatial tasks like gaze localization. To bridge this gap, we propose OmniGF, a unified vision-language framework that adapts foundational VLMs for highly scalable multi-person gaze reasoning. The model adopts a dual-branch decoding strategy: a structured language branch generates discrete reasoning states, while a continuous spatial branch directly taps into the VLM's dense hidden states. Supervising these extracted representations with high-resolution gaze target heatmaps effectively overcomes the spatial bottleneck of text-only coordinate generation. Furthermore, to explicitly ground the model in multi-person scenes, we augment the input with head embeddings encoded from cropped head images, providing fine-grained appearance and orientation cues for all individuals simultaneously. By modeling all individuals and leveraging the strong semantic capability of VLMs, OmniGF seamlessly integrates precise spatial gaze target estimation, semantic gaze prediction, and complex social gaze reasoning. Extensive experiments demonstrate that our framework establishes new state-of-the-art performance across multiple standard benchmarks. Code is available at https://github.com/cvlab-stonybrook/omnigf.

Lightweight vision-language models perform competitively on standard benchmarks yet fail systematically in dense-scene reasoning, where multiple objects, attributes, and relations must be jointly grounded and resolved through multi-step inference. Such capability is critical for real-world applications where models must reliably interpret cluttered environments. Yet existing training signals provide no explicit grounding between reasoning steps and the underlying visual entities and relations, leaving lightweight models free to generate fluent but visually unanchored reasoning chains. To address this gap, we first introduce DRBench, a benchmark of 14,573 questions across 2,943 images, organized into five task categories spanning three progressive reasoning layers. Building on DRBench, we propose DRScaffold, a supervised fine-tuning framework that decomposes the supervision target into four causally ordered stages, enforcing grounded reasoning without architectural modification. Experiments on three lightweight VLMs demonstrate substantial gains on DRBench while preserving or improving performance on general-purpose benchmarks. Notably, Qwen2.5-VL-3B trained with DRScaffold surpasses the frozen Qwen2.5-VL-32B on DRBench, demonstrating that structured supervision can substitute for a significant portion of model scale in dense-scene reasoning. Our code and models are available at https://github.com/irene-shi/DRScaffold .

Referring expression comprehension (REC) aims to localize a target object within an image based on a given expression. Although recent advances in vision-language models have led to substantial improvements in REC tasks, current REC benchmarks often hold simple scenarios and the assumption that each expression maps to a unique object. These limitations hinder the deployment of REC models in open-world environments. To fill this gap, we introduce OpenRef, a new benchmark for REC in complex visual and linguistic scenarios. OpenRef features three key advancements: 1) Diverse visual scenarios: spanning diverse visual domains, including ground views, drone views, dark scenes and adverse weather conditions; 2) Variable target counts: breaking the single-target limitation with multi-target and none-target samples; 3) Rich vocabulary types: incorporating proper nouns, polysemous words and ordinal terms to fit a wider range of expression needs. Furthermore, as traditional metrics are insufficient for open-world setting, we leverage F1 to measure grounding accuracy and propose N3R (Negative Relative Rejection Reliability) to assess relative rejection reliability against negative expressions. Finally, we introduce Multi-task Consistency Checker (MCC), a training-free but plug-and-play strategy that enhances model performance with one click by enforcing consistency self-verification. Extensive experiments demonstrate that this work significantly advances the performance of existing REC models in complex scenarios, paving the way for open-world REC. Project page: https://zongjianwu.github.io/openref

Despite success on standard benchmarks, vision language models display persistent failures on tasks involving processing of multi-object scenes, including many tasks that are relatively easy for humans. Recent work has found that these failures may stem from a basic inability to accurately bind object features in-context, a challenge that is referred to as the "binding problem" in cognitive science and neuroscience. The human visual system is thought to solve this binding problem via serial processing, attending to individual objects one at a time so as to avoid interference from other objects. Recent work has proposed "pointing" -- the use of explicit spatial coordinates to refer to objects -- as an analogous solution for vision language models, and found that it improves performance on challenging multi-object tasks. However, it is unclear $\textit{why}$ (i.e., on a mechanistic or representational level) this approach improves performance, and how directly this relates to serial processing in human vision. Here, we investigate this question. We find that learning to point-via-text induces an internal visual search routine, and we characterize the mechanisms that support this procedure. We also find that pointing behavior can be generalized to new tasks via fine-tuning, and that doing so eliminates binding errors and enables compositional generalization. These results provide a proof-of-principle that serial processing can solve the binding problem for vision language models just as it does for biological vision.

Spatial understanding of the physical world from 2D visual inputs hinges on two complementary forms of geometric knowledge: holistic 3D structural perception and fine-grained metric scale estimation. Existing multimodal large language models (MLLMs) typically address only one facet, ingesting either depth maps or point clouds as additional model inputs, which incurs substantial computational overhead and inherits the generalization limitations of upstream prediction models. We propose GAMSI, a dual-pathway Geometry-Aware MLLM for Spatial Intelligence that takes only RGB images as input while internalizing both forms of geometric prior within a unified autoregressive backbone. Specifically, we introduce Metric-Structure Decoupled Queries (MSDQ) which employ two groups of learnable queries to respectively extract dense metric signals and sparse structural cues from the shared visual context, with a task-decoupled attention mask further preventing the two pathways from contaminating each other. Building on this, an Expert-Guided Visual Grounding (EVG) module projects the aggregated cues back to frame-level visual features and aligns them with vision foundation models, which serve purely as training-time supervision, rather than as model inputs. We further build a multi-task spatial instruction-tuning dataset (MTS) comprising 152{,}776 samples spanning 13 task types and three visual modalities, consolidated from six public datasets. Trained with a two-stage curriculum, GAMSI achieves state-of-the-art performance on seven spatial intelligence benchmarks.

Despite remarkable progress in Multimodal Large Language Models (MLLMs), these models still struggle with fine-grained understanding tasks. In this work, we propose Procedurally Generated Tasks (PGT), a simple data-driven framework that serves a dual purpose: inducing fine-grained visual understanding and acting as a low-cost diagnostic tool to identify the source of perception failures. By overlaying unambiguous geometric primitives on images, PGT generate additional dense supervision that disentangles visual grounding capability from semantic priors. Extensive experiments on relational, quantitative, and 3D/depth understanding benchmarks show that PGT yields remarkable gains across diverse architectures. Instruction tuning MLLMs on LLaVA-v1.5-Instruct augmented with PGT data results in improvements of up to +20% on the What'sUp benchmark and +13.3% on CV-Bench-2D, while maintaining general perception capabilities. Moreover, finetuning state-of-the-art MLLMs on PGT data leads to boosts of up to +5.5% on What'sUp and +8.3% on CV-Bench-2D. These findings demonstrate that PGT effectively address the bottleneck of fine-grained perception, revealing that many spatial reasoning deficits stem from inadequate supervision signals rather than inherent architectural or resolution limitations.

Benchmark accuracy is often implicitly assumed to reflect grounded visual understanding in vision-language models (VLMs), yet it remains unclear to what extent such scores truly reflect reliance on visual evidence. Motivated by a surprising observation that removing a substantial fraction of image tokens only degrades model performance very slightly on a widely used hallucination benchmark, we systematically investigate this mismatch in a set of open-source VLMs. Our analysis spans multiple levels of granularity, spanning global visual degradation, localized occlusion, question reformulation, answer-space expansion, and decision-level analyses beyond standard accuracy. We further complement these behavioral results with a layer-wise analysis of vision-token geometry. Throughout the experiments, we find that although VLMs do incorporate visual input, their predictions are less sensitive to the loss of fine-grained visual evidence that standard accuracy should have suggested. Even when the final prediction remains unchanged, the model's internal support for the correct answer may already be weakened. We further complement a representation-level analysis, which shows increasing similarity among visual tokens in deeper layers, providing a possible explanation for our findings. Together, these results suggest that current benchmarks are not sufficient to reliably evaluate fine-grained visual grounding in VLMs.

Generative Vision-Language Models (VLMs) perform well on multimodal reasoning, but how visual inputs are transformed to text remains poorly understood. Existing interpretability work on VLMs uses Sparse Autoencoders (SAEs), which decompose static residual representations and miss the functional updates that drive cross-modal interaction. We adopt a function-centric framework based on Transcoders, sparse approximations of MLP sublayers that act as a causal proxy for layer-wise computation. Applied to Gemma 3-4B-IT, the framework decomposes the model into interpretable computational pathways linking image patches to directions in token generation. Transcoder attributions produce stronger and more stable effects on visually grounded tokens under patch ablation than SAE attributions, and align better with semantically relevant image regions. A False Visual Grounding counterfactual analysis confirms that the recovered pathways are specific to vision-language interaction.Finally, we perform a structural analysis of hallucinated generations, by extracting graph-based indicators from circuit traces produced by the transcoders. A logistic classifier over these mechanistic graph features predicts hallucinations at AUC $0.68$. These results show that function-centric circuit decomposition yields interpretable and predictive accounts of multimodal computation in VLMs.

Vision-language models (VLMs) often fail under low illumination because their visual grounding is learned predominantly from RGB imagery, whereas thermal infrared preserves complementary scene structure when visible cues degrade. We present Thermo-VL, a wavelength-aware VLM that augments a frozen Molmo-7B backbone with a trainable thermal encoder and a text-guided dual-attention fusion module. Given aligned RGB tokens, thermal tokens, and prompt embeddings, the fusion module conditions thermal features on both language and RGB context, then injects a gated residual into the frozen RGB stream so thermal evidence can be incorporated without disrupting Molmo's pretrained RGB-language interface. We train the model with the standard language-modeling objective together with auxiliary alignment and regularization losses that improve cross-modal grounding and reduce over-reliance on RGB. We also introduce a pixel-aligned RGB-thermal instruction-tuning dataset and Thermo-VL-Bench, a manually screened RGB-thermal VQA benchmark for low-light and cross-spectrum reasoning. Experiments show strong gains on challenging thermal-only and RGB+thermal reasoning tasks, highlighting the value of prompt-conditioned multispectral fusion. Our dataset and code are publicly available at: https://thusharakart.github.io/Thermo-VL

We introduce spatially grounded contextual image generation, a controllable image generation task that reframes the conditioning paradigm. Instead of supplying a reference image and a global text prompt through two separate encoders, one for vision and one for language, UniVL is trained to bind semantics to spatial locations directly from a single unified visual input, where the textual instruction is rendered onto the spatial mask. This removes the need for a standalone text encoder at inference time. The resulting model supports contextual image generation by following user-specified instructions about what should appear where, while substantially reducing computation. To address this task, we propose a framework in which the UniVL encoder, adapted from an optical-character-recognition-pretrained backbone, reads the unified condition optically and produces a UniVL embedding, fVIL, that fuses visual and semantic intent with spatial locations in a single token sequence. A two-stage pipeline first aligns UniVL with the VAE embedding space and then conditions a pretrained diffusion backbone entirely on UniVL embeddings, eliminating the standalone text encoder, such as T5. Although this reframing uses a deliberately minimal text interface, it yields strong empirical gains. On UniVL-ImgGen, a benchmark of 477K mask-annotated images that we construct for training and evaluation, UniVL improves image quality over text-prompted baselines, reducing FID from 14 to 11 and increasing PSNR from 16 to 20. It also eliminates the text encoder entirely, reducing inference TFLOPs by up to 52% and runtime by up to 44%. Additional ablation studies validate the contributions of the proposed components, paving the way for efficient, spatially grounded image generation with a unified conditioning paradigm.

Establishing a clear link between model predictions and the visual evidence that supports them is critical for transparency and reliability in multimodal reasoning, yet current multimodal large language model (MLLM) evaluations do not explicitly enforce this alignment. Existing benchmarks assess either textual answer correctness or pixel-level localization in isolation, leaving the coupling of reasoning and grounding an open challenge. We introduce VISTAQA, a comprehensive benchmark for joint evaluation of free-form answer correctness and pixel-level evidence grounding in visual question answering. VISTAQA comprises 1,157 expert-curated samples spanning six task types and six visual domains, ranging from direct perception to compositional and relational reasoning. VISTAQA requires models to not only answer correctly, but to also provide precise segmentation masks that support their answers. It also includes hallucination-aware examples where no valid visual evidence exists. To support this enhanced evaluation, we introduce GROVE, a unified evaluation metric that enforces joint correctness by combining textual accuracy and grounding quality via a per-sample geometric mean, ensuring neither dimension can compensate for deficiencies in the other. Comprehensive experiments across grounding-aware models and hybrid pipelines with general-purpose MLLMs reveal that even the strongest systems achieve limited performance under GROVE, highlighting a substantial gap between answer accuracy and visual evidence alignment.

We introduce WildRoadBench, a wild aerial road-damage grounding benchmark that couples direct visual grounding by vision-language models with autonomous research-and-engineering by LLM-driven agents on a single professionally annotated UAV corpus. The same image set and the same per-class AP_50 metric are evaluated under two protocols. The VLM Track measures whether a fixed VLM can localise domain-specific damage from one image and one short prompt under a unified prompting, decoding and parsing pipeline. The Agent Track measures whether an autonomous agent, given only a written task brief, a small exploratory slice and a fixed interaction budget, can search the public web, adapt pretrained components, write training and inference code, and submit predictions through a scalar-feedback oracle on a hidden holdout. We benchmark a broad pool of closed-source frontier models and open-source VLMs together with several frontier LLM-driven agents. Both routes remain far from reliable performance in this wild setting: closed-source frontier models lead the VLM leaderboard but still leave more than half of the metric on the table; open-source grounders plateau well below them, and newer generations or reasoning-style variants do not consistently improve grounding; small targets collapse for every open-source model; agents lag the strongest VLM despite richer affordances, and several fail to land a valid submission within the budget. We release the code and data at https://anonymous.4open.science/r/wildroadbench-0607 to support reproducible follow-up research.

Visually-grounded language models (VLMs) are highly effective in linking visual and textual information, yet they often struggle with basic classification and localization tasks. While classification mechanisms have been studied more extensively, the processes that support object localization remain poorly understood. In this work, we investigate two representative families, LLaVA-1.5 and InternVL-3.5, using a suite of mechanistic interpretability tools, including token ablations, attention knockout, and causal mediation analysis. We find that localization is driven by a containerization mechanism in which object-aligned tokens define the spatial extent of the object, while the semantic arrangement of tokens within those boundaries is largely irrelevant to the predicted box. Only a very small set of attention heads mediates the causal effect for both classification and localization, concentrating in early-mid layers for LLaVA and mid-late layers for InternVL. The two tasks share some early processing but ultimately depend on largely distinct specialized heads. Overall, we provide the first layer- and head-level account of localization in VLMs, revealing narrow computational pathways that can guide future model design and grounding objectives.

Vision-and-language navigation (VLN) requires an embodied agent to ground natural-language instructions into executable navigation actions in unseen environments. Existing zero-shot methods typically rely on additional waypoint prediction modules, which often entangle high-level directional reasoning with fine-grained local grounding, leading to error-prone and unstable decisions. In this paper, we propose P2DNav, a hierarchical framework for zero-shot vision-and-language navigation. P2DNav consists of three core components: Panorama-to-Downview (P2D), Sliding-Window Dialogue Memory (SDM), and Reflective Reorientation Mechanism (RRM). P2D explicitly decomposes navigation decision-making into two stages: panoramic direction selection and downview local grounding. It first selects the instruction-relevant direction from a 360° panorama, and then predicts a pixel-level target point from the downview RGB observation in that direction. In addition, SDM organizes navigation history as a multi-turn dialogue context and maintains recent visual observations within a sliding window to support long-horizon navigation. RRM further enables reflective reorientation by assessing the reliability of local grounding based on the downview observation and returning to panoramic direction selection when necessary. Experiments on the R2R-CE benchmark show that P2DNav achieves strong performance among zero-shot methods. In particular, compared with the state-of-the-art (SOTA) zero-shot waypoint-based and waypoint-free methods, P2DNav achieves SR gains of 146.6% and 58.9%, respectively, demonstrating the effectiveness of P2D, SDM, and RRM for zero-shot VLN. Code will be released for public use.

Children acquire language grounding with remarkable robustness from limited visuo-linguistic input in ways that surpass today's best large multimodal models. Recent research suggests current vision-language models (VLMs) trained on curated web data fail to generalize to the sparse, weakly-aligned egocentric streams produced by wearable devices, embodied agents, and infant head-cams -- and no fixed evaluation pipeline exists for measuring progress on this regime. We train VLMs on datasets with varying degrees of semantic alignment between visual and linguistic inputs, including naturalistic infant and adult egocentric videos, and evaluate them with a comprehensive suite spanning multimodal language grounding and unimodal vision and language tasks. At the core of this suite is Machine-DevBench, a corpus-grounded benchmark of lexical and grammatical competence, automatically generated from the model's training vocabulary across logarithmic frequency bins to eliminate the train/eval mismatch and low statistical power of prior developmental benchmarks. Our results show that current VLM paradigms hinge on the tight semantic alignment of curated data and fail to exploit the weakly-aligned signal that dominates naturalistic egocentric input -- the very regime in which humans thrive. To motivate progress, we introduce the EgoBabyVLM Challenge to drive the development of models capable of grounded language learning from the kind of naturalistic data that human infants experience.

Embodied control increasingly requires models to follow compositional language instructions while reasoning over dynamic visual states. However, current vision-language-action policies and world-action models often couple linguistic knowledge with visual computation in a shared backbone or conditioning pathway, leading to modality competition and making knowledge extension dependent on backbone updates. In this paper, we introduce Key-Gram, a conditional-memory framework that separates language-derived world knowledge from visual-state reasoning for embodied control. At its core is a memory module that decomposes an instruction into task-specific key-grams, retrieves static linguistic priors through deterministic hashed lookup, and injects the retrieved entries into selected hidden layers through context-aware gating and lightweight convolutional fusion. This design allows the backbone to devote its main capacity to visual reasoning and action inference, while reusable instruction knowledge is stored in an extensible external memory. The logical memory table can be conveniently partitioned during training and, due to its $O(1)$ lookup pattern, efficiently placed on host memory during inference. Across RoboTwin2.0, LIBERO/LIBERO-Plus, and real-world dual-arm manipulation, Key-Gram consistently improves both $π_{0}$ and $π_{0.5}$ backbones, with average relative gains of $29.5\%/9.9\%$ on RoboTwin2.0, $35.8\%/4.5\%$ on LIBERO-Plus transfer without target-domain fine-tuning, and $15.4\%/8.1\%$ on real-world long-horizon tasks. These results demonstrate that externalized linguistic memory provides an effective and extensible mechanism for improving compositional grounding, transfer, and real-world manipulation.

Leveraging the universal representations of pre-trained LLMs and MLLMs offers a promising path toward brain foundation models. However, visually-evoked EEG datasets remain scarce, leading existing methods to align neural signals mainly with abstract text, a lossy translation that may discard fine-grained perceptual information encoded in brain activity. We propose Generative Visual Grounding (GVG), a framework that visualizes the invisible by using an EEG-to-image generative model as a visual translator. Instead of forcing EEG into text alone, GVG hallucinates instance-specific proxy images for non-visual EEG, providing structured visual contexts that allow MLLMs to exploit their visual priors for clinical-state interpretation. We validate this idea on two MLLM backbones, GVG-X-Omni and GVG-Janus. Image-only alignment is already competitive: the lightweight GVG-X-Omni matches 1.7B-parameter text-aligned baselines while tuning only 170M parameters on a frozen 7B backbone. We further extend GVG-Janus with trimodal Image+Text alignment, where text supplies categorical semantic anchors and visual proxies enrich neural representations with perceptual details. Experiments show consistent gains in EEG understanding and visual generation, suggesting visual proxy grounding as an effective complement to textual alignment.

In recent years, multimodal large language models (MLLMs) have achieved remarkable progress, primarily attributed to effective paradigms for integrating visual and textual information. The dominant connector-based paradigm projects visual features into textual sequence, enabling unified multimodal alignment and reasoning within a generative architecture. However, our experiments reveal two key limitations: (1) Although visual information serves as the core evidential modality in MLLMs, it is treated on par with textual tokens, diminishing the unique contribution of the visual modality; (2) As generation length increases, particularly within a limited context window, the model's dependence on visual information progressively weakens, resulting in deteriorated vision-language alignment and reduced consistency between generated content and visual semantics. To address these challenges, we propose the Vision Inference Former (VIF), a lightweight architectural module that establishes a direct bridge between pure visual representations and the model's output space. Specifically, VIF continuously injects visual semantics throughout the decoding phase of the inference process, ensuring that the model remains firmly grounded in visual content during generation. We conduct experiments on 14 benchmark tasks covering general reasoning, OCR, table understanding, vision-centric evaluation, and hallucination. Experimental results show that VIF consistently improves model performance across diverse architectures while introducing minimal additional overhead. The code for this work is available at https://github.com/Dong-Xinpeng/VIF.

We present SWIM (See What I Mean), a novel training strategy that aligns vision and language representations to enable fine-grained object understanding solely from textual prompts. Unlike existing approaches that require explicit visual prompts, such as masks or points, SWIM leverages mask supervision only during training to guide cross-modal attention, allowing the model to automatically attend to the user-specified object at inference. Our cross-attention analysis of pretrained multimodal large languagemodels (MLLMs) reveals a systematic discrepancy: Attribute words produce sharp, localized activations in the visual modality, whereas object nouns yield diffuse and scattered patterns due to semantic reference bias and distributed high-level representations. To address this misalignment, we construct NL-Refer, an enriched dataset, in which each object mask is paired with a precise natural language referring expression. SWIM extracts multi-layer cross-attention maps from object nouns and enforces spatial consistency with ground-truth masks. Experimental results demonstrate that SWIM substantially improves text-visual alignment and achieves superior performance over visual-prompt-based methods on fine-grained object understanding benchmarks. The code and data are available at https://github.com/HumanMLLM/SWIM.

Remote sensing vision-language models commonly rely on pretrained visual encoders to convert images into semantic features before language-model reasoning. While effective for scene-level understanding, this pipeline may prematurely compress local visual evidence, making fine-grained spatial reasoning vulnerable to language priors, especially in ultra-high-resolution remote sensing imagery. We present SkyNative, a native multimodal framework for remote sensing that adopts an encoder-free architecture, removing the pretrained visual backbone to directly represent images as raw patch tokens in the language-model token space. To reconcile low-level visual patches with textual tokens, SkyNative introduces a modality-aware decoupling mechanism that uses modality-specific parameters within a unified autoregressive backbone. We further introduce a visual reliance benchmark that diagnoses whether models ground their answers in image evidence through progressive visual degradation and misleading textual prompts. Across standard remote sensing understanding tasks and large-format spatial reasoning evaluations, SkyNative shows stronger image-grounded perception and improved robustness against prompt-induced language priors. These results suggest that native patch-level multimodal modeling is a promising direction for reliable remote sensing vision-language reasoning.