KV-CoRE: Benchmarking Data-Dependent Low-Rank Compressibility of KV-Caches in LLMs

Large language models rely on kv-caches to avoid redundant computation during autoregressive decoding, but as context length grows, reading and writing the cache can quickly saturate GPU memory bandwidth. Recent work has explored KV-cache compression, yet most approaches neglect the data-dependent nature of kv-caches and their variation across layers. We introduce KV-CoRE KV-cache Compressibility by Rank Evaluation), an SVD-based method for quantifying the data-dependent low-rank compressibility of kv-caches. KV-CoRE computes the optimal low-rank approximation under the Frobenius norm and, being gradient-free and incremental, enables efficient dataset-level, layer-wise evaluation. Using this method, we analyze multiple models and datasets spanning five English domains and sixteen languages, uncovering systematic patterns that link compressibility to model architecture, training data, and language coverage. As part of this analysis, we employ the Normalized Effective Rank as a metric of compressibility and show that it correlates strongly with performance degradation under compression. Our study establishes a principled evaluation framework and the first large-scale benchmark of kv-cache compressibility in LLMs, offering insights for dynamic, data-aware compression and data-centric model development.

AesRec: A Dataset for Aesthetics-Aligned Clothing Outfit Recommendation

Clothing recommendation extends beyond merely generating personalized outfits; it serves as a crucial medium for aesthetic guidance. However, existing methods predominantly rely on user-item-outfit interaction behaviors while overlooking explicit representations of clothing aesthetics. To bridge this gap, we present the AesRec benchmark dataset featuring systematic quantitative aesthetic annotations, thereby enabling the development of aesthetics-aligned recommendation systems. Grounded in professional apparel quality standards and fashion aesthetic principles, we define a multidimensional set of indicators. At the item level, six dimensions are independently assessed: silhouette, chromaticity, materiality, craftsmanship, wearability, and item-level impression. Transitioning to the outfit level, the evaluation retains the first five core attributes while introducing stylistic synergy, visual harmony, and outfit-level impression as distinct metrics to capture the collective aesthetic impact. Given the increasing human-like proficiency of Vision-Language Models in multimodal understanding and interaction, we leverage them for large-scale aesthetic scoring. We conduct rigorous human-machine consistency validation on a fashion dataset, confirming the reliability of the generated ratings. Experimental results based on AesRec further demonstrate that integrating quantified aesthetic information into clothing recommendation models can provide aesthetic guidance for users while fulfilling their personalized requirements.

Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utility

Vision language models (VLMs) extend the reasoning capabilities of large language models (LLMs) to cross-modal settings, yet remain highly vulnerable to multimodal jailbreak attacks. Existing defenses predominantly rely on safety fine-tuning or aggressive token manipulations, incurring substantial training costs or significantly degrading utility. Recent research shows that LLMs inherently recognize unsafe content in text, and the incorporation of visual inputs in VLMs frequently dilutes risk-related signals. Motivated by this, we propose Risk Awareness Injection (RAI), a lightweight and training-free framework for safety calibration that restores LLM-like risk recognition by amplifying unsafe signals in VLMs. Specifically, RAI constructs an Unsafe Prototype Subspace from language embeddings and performs targeted modulation on selected high-risk visual tokens, explicitly activating safety-critical signals within the cross-modal feature space. This modulation restores the model's LLM-like ability to detect unsafe content from visual inputs, while preserving the semantic integrity of original tokens for cross-modal reasoning. Extensive experiments across multiple jailbreak and utility benchmarks demonstrate that RAI substantially reduces attack success rate without compromising task performance.

EventFlash: Towards Efficient MLLMs for Event-Based Vision

Event-based multimodal large language models (MLLMs) enable robust perception in high-speed and low-light scenarios, addressing key limitations of frame-based MLLMs. However, current event-based MLLMs often rely on dense image-like processing paradigms, overlooking the spatiotemporal sparsity of event streams and resulting in high computational cost. In this paper, we propose EventFlash, a novel and efficient MLLM to explore spatiotemporal token sparsification for reducing data redundancy and accelerating inference. Technically, we build EventMind, a large-scale and scene-diverse dataset with over 500k instruction sets, providing both short and long event stream sequences to support our curriculum training strategy. We then present an adaptive temporal window aggregation module for efficient temporal sampling, which adaptively compresses temporal tokens while retaining key temporal cues. Finally, a sparse density-guided attention module is designed to improve spatial token efficiency by selecting informative regions and suppressing empty or sparse areas. Experimental results show that EventFlash achieves a $12.4\times$ throughput improvement over the baseline (EventFlash-Zero) while maintaining comparable performance. It supports long-range event stream processing with up to 1,000 bins, significantly outperforming the 5-bin limit of EventGPT. We believe EventFlash serves as an efficient foundation model for event-based vision.

PRISM: Parametrically Refactoring Inference for Speculative Sampling Draft Models

Large Language Models (LLMs), constrained by their auto-regressive nature, suffer from slow decoding. Speculative decoding methods have emerged as a promising solution to accelerate LLM decoding, attracting attention from both systems and AI research communities. Recently, the pursuit of better draft quality has driven a trend toward parametrically larger draft models, which inevitably introduces substantial computational overhead. While existing work attempts to balance the trade-off between prediction accuracy and compute latency, we address this fundamental dilemma through architectural innovation. We propose PRISM, which disaggregates the computation of each predictive step across different parameter sets, refactoring the computational pathways of draft models to successfully decouple model capacity from inference cost. Through extensive experiments, we demonstrate that PRISM outperforms all existing draft architectures, achieving exceptional acceptance lengths while maintaining minimal draft latency for superior end-to-end speedup. We also re-examine scaling laws with PRISM, revealing that PRISM scales more effectively with expanding data volumes than other draft architectures. Through rigorous and fair comparison, we show that PRISM boosts the decoding throughput of an already highly optimized inference engine by more than 2.6x.

Learning to Decode Against Compositional Hallucination in Video Multimodal Large Language Models

Current research on video hallucination mitigation primarily focuses on isolated error types, leaving compositional hallucinations, arising from incorrect reasoning over multiple interacting spatial and temporal factors largely underexplored. We introduce OmniVCHall, a benchmark designed to systematically evaluate both isolated and compositional hallucinations in video multimodal large language models (VLLMs). OmniVCHall spans diverse video domains, introduces a novel camera-based hallucination type, and defines a fine-grained taxonomy, together with adversarial answer options (e.g., "All are correct" and "None of the above") to prevent shortcut reasoning. The evaluations of 39 representative VLLMs reveal that even advanced models (e.g., Qwen3-VL and GPT-5) exhibit substantial performance degradation. We propose TriCD, a contrastive decoding framework with a triple-pathway calibration mechanism. An adaptive perturbation controller dynamically selects distracting operations to construct negative video variants, while a saliency-guided enhancement module adaptively reinforces grounded token-wise visual evidences. These components are optimized via reinforcement learning to encourage precise decision-making under compositional hallucination settings. Experimental results show that TriCD consistently improves performance across two representative backbones, achieving an average accuracy improvement of over 10%. The data and code can be find at https://github.com/BMRETURN/OmniVCHall.

Diff-PC: Identity-preserving and 3D-aware Controllable Diffusion for Zero-shot Portrait Customization

Portrait customization (PC) has recently garnered significant attention due to its potential applications. However, existing PC methods lack precise identity (ID) preservation and face control. To address these tissues, we propose Diff-PC, a diffusion-based framework for zero-shot PC, which generates realistic portraits with high ID fidelity, specified facial attributes, and diverse backgrounds. Specifically, our approach employs the 3D face predictor to reconstruct the 3D-aware facial priors encompassing the reference ID, target expressions, and poses. To capture fine-grained face details, we design ID-Encoder that fuses local and global facial features. Subsequently, we devise ID-Ctrl using the 3D face to guide the alignment of ID features. We further introduce ID-Injector to enhance ID fidelity and facial controllability. Finally, training on our collected ID-centric dataset improves face similarity and text-to-image (T2I) alignment. Extensive experiments demonstrate that Diff-PC surpasses state-of-the-art methods in ID preservation, facial control, and T2I consistency. Furthermore, our method is compatible with multi-style foundation models.

ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents

Existing GUI agent models relying on coordinate-based one-step visual grounding struggle with generalizing to varying input resolutions and aspect ratios. Alternatives introduce coordinate-free strategies yet suffer from learning under severe data scarcity. To address the limitations, we propose ToolTok, a novel paradigm of multi-step pathfinding for GUI agents, where operations are modeled as a sequence of progressive tool usage. Specifically, we devise tools aligned with human interaction habits and represent each tool using learnable token embeddings. To enable efficient embedding learning under limited supervision, ToolTok introduces a semantic anchoring mechanism that grounds each tool with semantically related concepts as natural inductive bias. To further enable a pre-trained large language model to progressively acquire tool semantics, we construct an easy-to-hard curriculum consisting of three tasks: token definition question-answering, pure text-guided tool selection, and simplified visual pathfinding. Extensive experiments on multiple benchmarks show that ToolTok achieves superior performance among models of comparable scale (4B) and remains competitive with a substantially larger model (235B). Notably, these results are obtained using less than 1% of the training data required by other post-training approaches. In addition, ToolTok demonstrates strong generalization across unseen scenarios. Our training & inference code is open-source at https://github.com/ZephinueCode/ToolTok.

Rethinking LLM-as-a-Judge: Representation-as-a-Judge with Small Language Models via Semantic Capacity Asymmetry

Large language models (LLMs) are widely used as reference-free evaluators via prompting, but this "LLM-as-a-Judge" paradigm is costly, opaque, and sensitive to prompt design. In this work, we investigate whether smaller models can serve as efficient evaluators by leveraging internal representations instead of surface generation. We uncover a consistent empirical pattern: small LMs, despite with weak generative ability, encode rich evaluative signals in their hidden states. This motivates us to propose the Semantic Capacity Asymmetry Hypothesis: evaluation requires significantly less semantic capacity than generation and can be grounded in intermediate representations, suggesting that evaluation does not necessarily need to rely on large-scale generative models but can instead leverage latent features from smaller ones. Our findings motivate a paradigm shift from LLM-as-a-Judge to Representation-as-a-Judge, a decoding-free evaluation strategy that probes internal model structure rather than relying on prompted output. We instantiate this paradigm through INSPECTOR, a probing-based framework that predicts aspect-level evaluation scores from small model representations. Experiments on reasoning benchmarks (GSM8K, MATH, GPQA) show that INSPECTOR substantially outperforms prompting-based small LMs and closely approximates full LLM judges, while offering a more efficient, reliable, and interpretable alternative for scalable evaluation.

ProFlow: Zero-Shot Physics-Consistent Sampling via Proximal Flow Guidance

Inferring physical fields from sparse observations while strictly satisfying partial differential equations (PDEs) is a fundamental challenge in computational physics. Recently, deep generative models offer powerful data-driven priors for such inverse problems, yet existing methods struggle to enforce hard physical constraints without costly retraining or disrupting the learned generative prior. Consequently, there is a critical need for a sampling mechanism that can reconcile strict physical consistency and observational fidelity with the statistical structure of the pre-trained prior. To this end, we present ProFlow, a proximal guidance framework for zero-shot physics-consistent sampling, defined as inferring solutions from sparse observations using a fixed generative prior without task-specific retraining. The algorithm employs a rigorous two-step scheme that alternates between: (\romannumeral1) a terminal optimization step, which projects the flow prediction onto the intersection of the physically and observationally consistent sets via proximal minimization; and (\romannumeral2) an interpolation step, which maps the refined state back to the generative trajectory to maintain consistency with the learned flow probability path. This procedure admits a Bayesian interpretation as a sequence of local maximum a posteriori (MAP) updates. Comprehensive benchmarks on Poisson, Helmholtz, Darcy, and viscous Burgers' equations demonstrate that ProFlow achieves superior physical and observational consistency, as well as more accurate distributional statistics, compared to state-of-the-art diffusion- and flow-based baselines.