We present a hybrid surrogate model for electric vehicle parameter estimation and power consumption. We combine our novel architecture Spectral Parameter Operator built on a Fourier Neural Operator backbone for global context and a differentiable physics module in the forward pass. From speed and acceleration alone, it outputs time-varying motor and regenerative braking efficiencies, as well as aerodynamic drag, rolling resistance, effective mass, and auxiliary power. These parameters drive a physics-embedded estimate of battery power, eliminating any separate physics-residual loss. The modular design lets representations converge to physically meaningful parameters that reflect the current state and condition of the vehicle. We evaluate on real-world logs from a Tesla Model 3, Tesla Model S, and the Kia EV9. The surrogate achieves a mean absolute error of 0.2kW (about 1% of average traction power at highway speeds) for Tesla vehicles and about 0.8kW on the Kia EV9. The framework is interpretable, and it generalizes well to unseen conditions, and sampling rates, making it practical for path optimization, eco-routing, on-board diagnostics, and prognostics health management.

Sentence embeddings are essential for NLP tasks such as semantic search, re-ranking, and textual similarity. Although multilingual benchmarks like MMTEB broaden coverage, Southeast Asia (SEA) datasets are scarce and often machine-translated, missing native linguistic properties. With nearly 700 million speakers, the SEA region lacks a region-specific embedding benchmark. We introduce SEA-BED, the first large-scale SEA embedding benchmark with 169 datasets across 9 tasks and 10 languages, where 71% are formulated by humans, not machine generation or translation. We address three research questions: (1) which SEA languages and tasks are challenging, (2) whether SEA languages show unique performance gaps globally, and (3) how human vs. machine translations affect evaluation. We evaluate 17 embedding models across six studies, analyzing task and language challenges, cross-benchmark comparisons, and translation trade-offs. Results show sharp ranking shifts, inconsistent model performance among SEA languages, and the importance of human-curated datasets for low-resource languages like Burmese.

--We study how gating mechanisms in recurrent neural networks (RNNs) implicitly induce adaptive learning-rate behavior, even when training is carried out with a fixed, global learning rate. This effect arises from the coupling between state-space time scales-parametrized by the gates-and parameter-space dynamics during gradient descent. By deriving exact Jacobians for leaky-integrator and gated RNNs, we obtain a first-order expansion that makes explicit how constant, scalar, and multi-dimensional gates reshape gradient propagation, modulate effective step sizes, and introduce anisotropy in parameter updates. These findings reveal that gates not only control information flow, but also act as data-driven preconditioners that adapt optimization trajectories in parameter space. Empirical simulations corroborate these claims: in several sequence tasks, we show that gates induce lag-dependent effective learning rates and directional concentration of gradient flow, with multi-gate models matching or exceeding the anisotropic structure produced by Adam. These results highlight that optimizer-driven and gate-driven adaptivity are complementary but not equivalent mechanisms. Overall, this work provides a unified dynamical systems perspective on how gating couples state evolution with parameter updates, explaining why gated architectures achieve robust trainability and stability in practice.

Despite extensive safety-tuning, large language models (LLMs) remain vulnerable to jailbreak attacks via adversarially crafted instructions, reflecting a persistent trade-off between safety and task performance. We comprehensively evaluate both parametric and non-parametric attacks across open-source and proprietary models, considering harmfulness from attacks, utility, over-refusal, and impact against white-box threats. Importantly, our method preserves the model's general capabilities and reduces excessive refusals on benign instructions containing superficially harmful keywords. With the rapid advancement of large language models (LLMs) (Grattafiori et al., 2024; Y ang et al., 2025; Liu et al., 2024; Mao et al., 2024), the risk of these models executing harmful or catastrophic instructions has grown correspondingly (Anthropic, 2025). This is largely managed by efforts such as dedicated a safety-alignment stage (Ouyang et al., 2022), aiming to ensure that LLMs are not only helpful but also consistently generate safe and ethical outputs. Nevertheless, recent findings by Qi et al. (2024) expose a fundamental vulnerability in prevailing safety-alignment practices: Shallow Alignment . In particular, alignment in most models is largely superficial--constrained to surface-level refusals--resulting in safe outputs that are often limited to generic templates such as "I am sorry but... " or "As a language model... " . This superficial alignment permits attackers to circumvent safety mechanisms by explicitly instructing the model to avoid generating commonly recognized refusal responses (Tang, 2024; Andriushchenko et al., 2025). Furthermore, LLMs remain susceptible to a broader range of prompt-based attacks, including those that optimize over discrete suffix tokens (Zou et al., 2023; Basani & Zhang, 2025) or rephrase harmful instructions to look harmless (Chao et al., 2025; Zeng et al., 2024). Beyond initial safety alignment, practitioners have developed a range of inference-time defenses, such as prompting models to adhere to their safety guidelines (Xie et al., 2023) incorporating additional safety exemplars to enable in-context defense (Wei et al., 2023). Wang et al. (2024) introduce a backdoor trigger into safety-aligned LLMs, serving as a covert prefix that elicits safety responses when detected, without affecting model behavior on benign queries. In recent works, Zhang et al. (2024) introduced a dual-stage prompting strategy, Intention Analysis (IA), which encourages LLMs to analyze the intent behind an instruction prior to generating a safe response.

LLM-based agents have emerged as transformative tools capable of executing complex tasks through iterative planning and action, achieving significant advancements in understanding and addressing user needs. Yet, their effectiveness remains limited in specialized domains such as mental health diagnosis, where they underperform compared to general applications. Current approaches to integrating diagnostic capabilities into LLMs rely on scarce, highly sensitive mental health datasets, which are challenging to acquire. These methods also fail to emulate clinicians' proactive inquiry skills, lack multi-turn conversational comprehension, and struggle to align outputs with expert clinical reasoning. To address these gaps, we propose DSM5AgentFlow, the first LLM-based agent workflow designed to autonomously generate DSM-5 Level-1 diagnostic questionnaires. By simulating therapist-client dialogues with specific client profiles, the framework delivers transparent, step-by-step disorder predictions, producing explainable and trustworthy results. This workflow serves as a complementary tool for mental health diagnosis, ensuring adherence to ethical and legal standards. Through comprehensive experiments, we evaluate leading LLMs across three critical dimensions: conversational realism, diagnostic accuracy, and explainability. Our datasets and implementations are fully open-sourced.

We propose Frequency-Guided Attention (FGA), a lightweight upsampling module for single image super-resolution. Conventional upsamplers, such as Sub-Pixel Convolution, are efficient but frequently fail to reconstruct high-frequency details and introduce aliasing artifacts. FGA addresses these issues by integrating (1) a Fourier feature-based Multi-Layer Perceptron (MLP) for positional frequency encoding, (2) a cross-resolution Correlation Attention Layer for adaptive spatial alignment, and (3) a frequency-domain L1 loss for spectral fidelity supervision. Adding merely 0.3M parameters, FGA consistently enhances performance across five diverse super-resolution backbones in both lightweight and full-capacity scenarios. Experimental results demonstrate average PSNR gains of 0.12~0.14 dB and improved frequency-domain consistency by up to 29%, particularly evident on texture-rich datasets. Visual and spectral evaluations confirm FGA's effectiveness in reducing aliasing and preserving fine details, establishing it as a practical, scalable alternative to traditional upsampling methods.

Large language models (LLMs) excel at solving problems with clear and complete statements, but often struggle with nuanced environments or interactive tasks which are common in most real-world scenarios. This highlights the critical need for developing LLMs that can effectively engage in logically consistent multi-turn dialogue, seek information and reason with incomplete data. To this end, we introduce a novel benchmark comprising a suite of multi-turn tasks each designed to test specific reasoning, interactive dialogue, and information-seeking abilities. These tasks have deterministic scoring mechanisms, thus eliminating the need for human intervention. Evaluating frontier models on our benchmark reveals significant headroom. Our analysis shows that most errors emerge from poor instruction following, reasoning failures, and poor planning. This benchmark provides valuable insights into the strengths and weaknesses of current LLMs in handling complex, interactive scenarios and offers a robust platform for future research aimed at improving these critical capabilities.

Fine-tuning as service injects domain-specific knowledge into large language models (LLMs), while challenging the original alignment mechanisms and introducing safety risks. A series of defense strategies have been proposed for the alignment, fine-tuning, and post-fine-tuning phases, where most post-fine-tuning defenses rely on coarse-grained safety layer mapping. These methods lack a comprehensive consideration of both safety layers and fine-grained neurons, limiting their ability to efficiently balance safety and utility. To address this, we propose the Fine-Grained Safety Neurons (FGSN) with Training-Free Continual Projection method to reduce the fine-tuning safety risks. FGSN inherently integrates the multi-scale interactions between safety layers and neurons, localizing sparser and more precise fine-grained safety neurons while minimizing interference with downstream task neurons. We then project the safety neuron parameters onto safety directions, improving model safety while aligning more closely with human preferences. Extensive experiments across multiple fine-tuned LLM models demonstrate that our method significantly reduce harmfulness scores and attack success rates with minimal parameter modifications, while preserving the model's utility. Furthermore, by introducing a task-specific, multi-dimensional heterogeneous safety neuron cluster optimization mechanism, we achieve continual defense and generalization capability against unforeseen emerging safety concerns.

Solving the problem of Optical Character Recognition (OCR) on printed text for Latin and its derivative scripts can now be considered settled due to the volumes of research done on English and other High-Resourced Languages (HRL). However, for Low-Resourced Languages (LRL) that use unique scripts, it remains an open problem. This study presents a comparative analysis of the zero-shot performance of six distinct OCR engines on two LRLs: Sinhala and Tamil. The selected engines include both commercial and open-source systems, aiming to evaluate the strengths of each category. The Cloud Vision API, Surya, Document AI, and Tesseract were evaluated for both Sinhala and Tamil, while Subasa OCR and EasyOCR were examined for only one language due to their limitations. The performance of these systems was rigorously analysed using five measurement techniques to assess accuracy at both the character and word levels. According to the findings, Surya delivered the best performance for Sinhala across all metrics, with a WER of 2.61%. Conversely, Document AI excelled across all metrics for Tamil, highlighted by a very low CER of 0.78%. In addition to the above analysis, we also introduce a novel synthetic Tamil OCR benchmarking dataset.

This paper addresses the challenge of automated sports video analysis, which has traditionally been limited by computationally intensive models requiring server-side processing and lacking fine-grained understanding of athletic movements. Current approaches struggle to capture the nuanced biomechanical transitions essential for meaningful sports analysis, often missing critical phases like preparation, execution, and follow-through that occur within seconds. To address these limitations, we introduce SV3.3B, a lightweight 3.3B parameter video understanding model that combines novel temporal motion difference sampling with self-supervised learning for efficient on-device deployment. Our approach employs a DWT-VGG16-LDA based keyframe extraction mechanism that intelligently identifies the 16 most representative frames from sports sequences, followed by a V-DWT-JEPA2 encoder pretrained through mask-denoising objectives and an LLM decoder fine-tuned for sports action description generation. Evaluated on a subset of the NSVA basketball dataset, SV3.3B achieves superior performance across both traditional text generation metrics and sports-specific evaluation criteria, outperforming larger closed-source models including GPT-4o variants while maintaining significantly lower computational requirements. Our model demonstrates exceptional capability in generating technically detailed and analytically rich sports descriptions, achieving 29.2% improvement over GPT-4o in ground truth validation metrics, with substantial improvements in information density, action complexity, and measurement precision metrics essential for comprehensive athletic analysis. Model Available at https://huggingface.co/sportsvision/SV3.3B.