Adapting advertisements for multilingual audiences requires more than simple text translation; it demands preservation of visual consistency, spatial alignment, and stylistic integrity across diverse languages and formats. W e introduce a structured framework that combines automated components with human oversight to address the complexities of advertisement localization. T o the best of our knowledge, this is the first work to integrate scene text detection, inpainting, machine translation (MT), and text reimposition specifically for accelerating ad localization evaluation workflows. Qualitative results across six locales demonstrate that our approach produces semantically accurate and visually coherent localized advertisements, suitable for deployment in real-world workflows.

Prompt engineering is crucial for fully leveraging large language models (LLMs), yet most existing optimization methods follow a single trajectory, resulting in limited adaptability, gradient conflicts, and high computational overhead. We propose MAPGD (Multi-Agent Prompt Gradient Descent), a novel framework that reconceptualizes prompt optimization as a collaborative process among specialized agents. Each agent focuses on a distinct refinement dimension, such as instruction clarity, example selection, format structure, or stylistic adaptation, and their contributions are coordinated through semantic gradient embedding, conflict detection, and fusion. To further enhance robustness and stability, MAPGD introduces two new mechanisms: Hypersphere Constrained Gradient Clustering (HCGC), which enforces angular margin constraints for compact and well-separated clusters, and Channel Adaptive Agent Weighting (CAAW), which dynamically reweights agent contributions based on validation performance. Experiments on classification and reasoning benchmarks show that MAPGD consistently surpasses single-agent and random baselines in both accuracy and efficiency. Ablation studies confirm the effectiveness of gradient fusion, agent specialization, and conflict resolution. Together, these components establish MAPGD as a unified, gradient-based, and interpretable framework for robust prompt optimization with theoretical convergence guarantees.

Effective model and hyperparameter selection remains a major challenge in deep learning, often requiring extensive expertise and computation. While AutoML and large language models (LLMs) promise automation, current LLM-based approaches rely on trial and error and expensive APIs, which provide limited interpretability and generalizability. We propose MetaLLMiX, a zero-shot hyperparameter optimization framework combining meta-learning, explainable AI, and efficient LLM reasoning. By leveraging historical experiment outcomes with SHAP explanations, MetaLLMiX recommends optimal hyperparameters and pretrained models without additional trials. We further employ an LLM-as-judge evaluation to control output format, accuracy, and completeness. Experiments on eight medical imaging datasets using nine open-source lightweight LLMs show that MetaLLMiX achieves competitive or superior performance to traditional HPO methods while drastically reducing computational cost. Our local deployment outperforms prior API-based approaches, achieving optimal results on 5 of 8 tasks, response time reductions of 99.6-99.9%, and the fastest training times on 6 datasets (2.4-15.7x faster), maintaining accuracy within 1-5% of best-performing baselines.

Over half of US adults with Alzheimer disease and related dementias remain undiagnosed, and speech-based screening offers a scalable detection approach. We compared large language model adaptation strategies for dementia detection using the DementiaBank speech corpus, evaluating nine text-only models and three multimodal audio-text models on recordings from DementiaBank speech corpus. Adaptations included in-context learning with different demonstration selection policies, reasoning-augmented prompting, parameter-efficient fine-tuning, and multimodal integration. Results showed that class-centroid demonstrations achieved the highest in-context learning performance, reasoning improved smaller models, and token-level fine-tuning generally produced the best scores. Adding a classification head substantially improved underperforming models. Among multimodal models, fine-tuned audio-text systems performed well but did not surpass the top text-only models. These findings highlight that model adaptation strategies, including demonstration selection, reasoning design, and tuning method, critically influence speech-based dementia detection, and that properly adapted open-weight models can match or exceed commercial systems.

Large language models (LLMs) are increasingly seen as assistants, copilots, and consultants, capable of supporting a wide range of tasks through natural conversation. However, most systems remain constrained by a linear request-response format that often makes interactions inefficient in multi-turn, information-dense, and exploratory tasks. To address these limitations, we propose Generative Interfaces for Language Models, a paradigm in which LLMs respond to user queries by proactively generating user interfaces (UIs) that enable more adaptive and interactive engagement. Our framework leverages structured interface-specific representations and iterative refinements to translate user queries into task-specific UIs. For systematic evaluation, we introduce a multidimensional assessment framework that compares generative interfaces with traditional chat-based ones across diverse tasks, interaction patterns, and query types, capturing functional, interactive, and emotional aspects of user experience. Results show that generative interfaces consistently outperform conversational ones, with up to a 72% improvement in human preference. These findings clarify when and why users favor generative interfaces, paving the way for future advancements in human-AI interaction.

In this paper, we study the surprising impact that truncating text embeddings has on downstream performance. We consistently observe across 6 state-of-the-art text encoders and 26 downstream tasks, that randomly removing up to 50% of embedding dimensions results in only a minor drop in performance, less than 10%, in retrieval and classification tasks. Given the benefits of using smaller-sized embeddings, as well as the potential insights about text encoding, we study this phenomenon and find that, contrary to what is suggested in prior work, this is not the result of an ineffective use of representation space. Instead, we find that a large number of uniformly distributed dimensions actually cause an increase in performance when removed. This would explain why, on average, removing a large number of embedding dimensions results in a marginal drop in performance. We make similar observations when truncating the embeddings used by large language models to make next-token predictions on generative tasks, suggesting that this phenomenon is not isolated to classification or retrieval tasks.

Abstract--AI-based solutions demonstrate remarkable results in identifying vulnerabilities in software, but research has consistently found that this performance does not generalize to unseen codebases. In this paper, we specifically investigate the impact of model architecture, parameter configuration, and quality of training data on the ability of these systems to generalize. For this purpose, we introduce V ulGate, a high quality state of the art dataset that mitigates the shortcomings of prior datasets, by removing mislabeled and duplicate samples, updating new vulnerabilities, incorporating additional metadata, integrating hard samples, and including dedicated test sets. We undertake a series of experiments to demonstrate that improved dataset diversity and quality substantially enhances vulnerability detection. We also introduce and benchmark multiple encoder-only and decoder-only models. We find that encoder-based models outperform other models in terms of accuracy and generalization. Our model achieves 6.8% improvement in recall on the benchmark BigV ul dataset and outperforms others on unseen projects, demonstrating enhanced generalizability. Our results highlight the role of data quality and model selection in the development of robust vulnerability detection systems. Our findings suggest a direction for future systems with high cross-project effectiveness. With the rapid growth in digitization and software applications and systems in recent years, the issue of software vulnerabilities has become a critical concern. In 2024, a record-breaking 40,000 Common Vulnerabilities and Exposures (CVEs) were published--an average of 108 per day--marking a 38% increase over 2023 (with 28,818 CVEs) [1]. This number is already dramatically increasing: the first half of 2025 has witnessed an average of 131 CVEs per day [2]. In the open-source software ecosystem, which underpins a wide range of industries, including finance, energy, aerospace, and healthcare, a recent study found a surge of 98% per year in reported vulnerabilities [3]. R. Safdar, S.T. Ali and W . Hussain are with School of Electrical Engineering and Computer Science, National University of Sciences and Technology, Islamabad, Pakistan, 44000.

Ensuring the safe use of agentic systems requires a thorough understanding of the range of malicious behaviors these systems may exhibit when under attack. In this paper, we evaluate the robustness of LLM-based agentic systems against attacks that aim to elicit harmful actions from agents. To this end, we propose a novel taxonomy of harms for agentic systems and a novel benchmark, BAD-ACTS, for studying the security of agentic systems with respect to a wide range of harmful actions. BAD-ACTS consists of 4 implementations of agentic systems in distinct application environments, as well as a dataset of 188 high-quality examples of harmful actions. This enables a comprehensive study of the robustness of agentic systems across a wide range of categories of harmful behaviors, available tools, and inter-agent communication structures. Using this benchmark, we analyze the robustness of agentic systems against an attacker that controls one of the agents in the system and aims to manipulate other agents to execute a harmful target action. Our results show that the attack has a high success rate, demonstrating that even a single adversarial agent within the system can have a significant impact on the security. This attack remains effective even when agents use a simple prompting-based defense strategy. However, we additionally propose a more effective defense based on message monitoring. We believe that this benchmark provides a diverse testbed for the security research of agentic systems. The benchmark can be found at github.com/JNoether/BAD-ACTS

Large language models (LLMs) have demonstrated significant capabilities in solving mathematical problems expressed in natural language. However, multilingual and culturally-grounded mathematical reasoning in low-resource languages lags behind English due to the scarcity of socio-cultural task datasets that reflect accurate native entities such as person names, organization names, and currencies. Existing multilingual benchmarks are predominantly produced via translation and typically retain English-centric entities, owing to the high cost associated with human annotater-based localization. Moreover, automated localization tools are limited, and hence, truly localized datasets remain scarce. To bridge this gap, we introduce a framework for LLM-driven cultural localization of math word problems that automatically constructs datasets with native names, organizations, and currencies from existing sources. We find that translated benchmarks can obscure true multilingual math ability under appropriate socio-cultural contexts. Through extensive experiments, we also show that our framework can help mitigate English-centric entity bias and improves robustness when native entities are introduced across various languages.

Modern large language models often encode sensitive, harmful, or copyrighted knowledge, raising the need for post-hoc unlearning-the ability to remove specific domains of knowledge from a model without full retraining. A major bottleneck in current unlearning pipelines is constructing effective forget sets-datasets that approximate the target domain and guide the model to forget it. In this work, we introduce a scalable, automated approach to generate high-quality forget sets using language models themselves. Our method synthesizes textbook-style data through a structured prompting pipeline, requiring only a domain name as input. Through experiments on unlearning biosecurity, cybersecurity, and Harry Potter novels, we show that our synthetic datasets consistently outperform the baseline synthetic alternatives and are comparable to the expert-curated ones. Additionally, ablation studies reveal that the multi-step generation pipeline significantly boosts data diversity, which in turn improves unlearning utility. Overall, our findings suggest that synthetic datasets offer a promising path toward practical, scalable unlearning for a wide range of emerging domains without the need for manual intervention. We release our code and dataset at https://github.com/xyzhu123/Synthetic_Textbook.