Large Multimodal Models have achieved remarkable progress in integrating vision and language, enabling strong performance across perception, reasoning, and domain-specific tasks. However, their capacity to reason over multiple, visually similar inputs remains insufficiently explored. Such fine-grained comparative reasoning is central to real-world tasks, especially in mathematics and education, where learners must often distinguish between nearly identical diagrams to identify correct solutions. To address this gap, we present VisioMath, a curated benchmark of 1,800 high-quality K-12 mathematics problems in which all candidate answers are diagrams with subtle visual similarities. A comprehensive evaluation of state-of-the-art LMMs, covering both leading closed-source systems and widely adopted open-source models, reveals a consistent decline in accuracy as inter-image similarity increases. Analysis indicates that the dominant failure mode stems from image-text misalignment: rather than grounding reasoning in textual cues, models often resort to shallow positional heuristics, resulting in systematic errors. We further explore three alignment-oriented strategies, spanning training-free approaches and finetuning, and achieve substantial accuracy gains. We hope that VisioMath will serve as a rigorous benchmark and catalyst for developing LMMs toward deeper diagram understanding, precise comparative reasoning, and grounded multi-image-text integration.

Recent advances in large language models (LLMs) have sparked growing interest in integrating language-driven techniques into trajectory prediction. By leveraging their semantic and reasoning capabilities, LLMs are reshaping how autonomous systems perceive, model, and predict trajectories. This survey provides a comprehensive overview of this emerging field, categorizing recent work into five directions: (1) Trajectory prediction via language modeling paradigms, (2) Direct trajectory prediction with pretrained language models, (3) Language-guided scene understanding for trajectory prediction, (4) Language-driven data generation for trajectory prediction, (5) Language-based reasoning and interpretability for trajectory prediction. For each, we analyze representative methods, highlight core design choices, and identify open challenges. This survey bridges natural language processing and trajectory prediction, offering a unified perspective on how language can enrich trajectory prediction.

Federated instruction tuning of large language models (LLMs) is challenged by significant data heterogeneity across clients, demanding robust personalization. The Mixture of Experts (MoE) architecture, where experts can specialize in distinct data patterns, presents a natural architectural solution to this challenge. The inherent sparsity of the MoE architecture, achieved by selectively activating experts, poses a significant challenge to its integration with federated learning (FL). Conventional FL frameworks, designed for dense models, naively aggregate all expert parameters irrespective of their local activation patterns. This naive approach not only undermines MoE's dynamic sparsity but also risks corrupting the world knowledge within pretrained experts. To address this, we propose FLEx (Federated LLMs with Personalized Experts), a novel framework that leverages pretrained MoE-based LLMs for efficient personalization. By aggregating only the shared non-expert parameters, FLEx significantly reduces communication overhead and preserves the world knowledge stored within the frozen pretrained experts. For personalization, we introduce a novel expert grafting mechanism that leverages dynamic sparsity to construct a client-specific expert from selected components of pretrained experts, tailored to local data. This grafted expert is then fine-tuned locally alongside the gating mechanism. This joint training enables the model to learn when to leverage the shared knowledge from frozen experts and when to employ the personalized one. Evaluations on diverse, non-IID instruction tuning datasets show that FLEx consistently outperforms federated baselines on average, while demonstrating strong knowledge preservation on the knowledge-driven benchmark MMLU. Our code is available at \href{https://anonymous.4open.science/r/FLEx-8F12}{\texttt{https://anonymous.4open.science/r/FLEx-8F12}}.

Large language models (LLMs) are known to memorize and recall English text from their pretraining data. However, the extent to which this ability generalizes to non-English languages or transfers across languages remains unclear. This paper investigates multilingual and cross-lingual memorization in LLMs, probing if memorized content in one language (e.g., English) can be recalled when presented in translation. To do so, we introduce OWL, a dataset of 31.5K aligned excerpts from 20 books in ten languages, including English originals, official translations (Vietnamese, Spanish, Turkish), and new translations in six low-resource languages (Sesotho, Yoruba, Maithili, Malagasy, Setswana, Tahitian). We evaluate memorization across model families and sizes through three tasks: (1) direct probing, which asks the model to identify a book's title and author; (2) name cloze, which requires predicting masked character names; and (3) prefix probing, which involves generating continuations. We find that LLMs consistently recall content across languages, even for texts without direct translation in pretraining data. GPT-4o, for example, identifies authors and titles 69% of the time and masked entities 6% of the time in newly translated excerpts. Perturbations (e.g., masking characters, shuffling words) modestly reduce direct probing accuracy (7% drop for shuffled official translations). Our results highlight the extent of cross-lingual memorization and provide insights on the differences between the models.

Inference scaling helps LLMs solve complex reasoning problems through extended runtime computation. On top of long chain-of-thought (long-CoT) models, purely inference-time techniques such as best-of-N (BoN) sampling, majority voting, or more generally, minimum Bayes risk decoding (MBRD), can further improve LLM accuracy by generating multiple candidate solutions and aggregating over them. These methods typically leverage additional signals in the form of reward models and risk/similarity functions that compare generated samples, e.g., exact match in some normalized space or standard similarity metrics such as Rouge. Here we present a novel method for incorporating reward and risk/similarity signals into MBRD. Based on the concept of optimal policy in KL-controlled reinforcement learning, our framework provides a simple and well-defined mechanism for leveraging such signals, offering several advantages over traditional inference-time methods: higher robustness, improved accuracy, and well-understood asymptotic behavior. In addition, it allows for the development of a sample-efficient variant of MBRD that can adjust the number of samples to generate according to the difficulty of the problem, without relying on majority vote counts. We empirically demonstrate the advantages of our approach on math (MATH-$500$) and coding (HumanEval) tasks using recent open-source models. We also present a comprehensive analysis of its accuracy-compute trade-offs.

The growing collaboration between humans and AI models in generative tasks has introduced new challenges in distinguishing between human-written, LLM-generated, and human--LLM collaborative texts. In this work, we collect a multilingual, multi-domain, multi-generator dataset FAIDSet. We further introduce a fine-grained detection framework FAID to classify text into these three categories, and also to identify the underlying LLM family of the generator. Unlike existing binary classifiers, FAID is built to capture both authorship and model-specific characteristics. Our method combines multi-level contrastive learning with multi-task auxiliary classification to learn subtle stylistic cues. By modeling LLM families as distinct stylistic entities, we incorporate an adaptation to address distributional shifts without retraining for unseen data. Our experimental results demonstrate that FAID outperforms several baselines, particularly enhancing the generalization accuracy on unseen domains and new LLMs, thus offering a potential solution for improving transparency and accountability in AI-assisted writing.

Reasoning has emerged as the next major frontier for language models (LMs), with rapid advances from both academic and industrial labs. However, this progress often outpaces methodological rigor, with many evaluations relying on benchmarking practices that lack transparency, robustness, or statistical grounding. In this work, we conduct a comprehensive empirical study and find that current mathematical reasoning benchmarks are highly sensitive to subtle implementation choices--including decoding parameters, random seeds, prompt formatting, and even hardware and software configurations. Performance gains reported in recent studies frequently hinge on unclear comparisons or unreported sources of variance. To address these issues, we propose a standardized evaluation framework with clearly defined best practices and reporting standards. Using this framework, we reassess recent methods and find that most reinforcement learning (RL) approaches yield only modest improvements--far below prior claims--and are prone to overfitting, especially on small-scale benchmarks like AIME'24. In contrast, supervised finetuning (SFT) methods show consistently stronger generalization in the settings we study. To foster reproducibility, we release all code, prompts, and model outputs, for reasoning benchmarks, establishing more rigorous foundations for future work.

AI toys are all the rage in China--and now they're appearing on shelves in the US too Competition is heating up, with Mattel and OpenAI expected to launch a product for kids this year. Kids have always played with and talked to stuffed animals. But now their toys can talk back, thanks to a wave of companies that are fitting children's playthings with chatbots and voice assistants. It's a trend that has particularly taken off in China: A recent report by the Shenzhen Toy Industry Association and JD.com predicts that the sector will surpass ¥100 billion ($14 billion) by 2030, growing faster than almost any other branch of consumer AI. According to the Chinese corporation registration database Qichamao, there are over 1,500 AI toy companies operating in China as of October 2025. One of the latest entrants to the market is a toy called BubblePal, a device the size of a Ping-Pong ball that clips onto a child's favorite stuffed animal and makes it "talk."

OpenAI revealed last week the custom AI tools it uses internally. The news sent some software companies into turmoil. Allan Thygesen, the CEO of Docusign, was not particularly concerned when he saw the news last week that OpenAI had created an internal tool called DocuGPT . He might have preferred that OpenAI choose a different name for its contracting tool. But still, he thought, DocuGPT barely scratched the surface of what Docusign can do.

In this still from the Sora 2 promotional video, an ai-generated cameo of Sam Altman shows us through worlds of generated content. Last week OpenAI released Sora, a TikTok-style app that presents an endless feed of exclusively AI-generated videos, each up to 10 seconds long. The app allows you to create a "cameo" of yourself--a hyperrealistic avatar that mimics your appearance and voice--and insert other peoples' cameos into your own videos (depending on what permissions they set). To some people who believed earnestly in OpenAI's promise to build AI that benefits all of humanity, the app is a punchline. A former OpenAI researcher who left to build an AI-for-science startup referred to Sora as an "infinite AI tiktok slop machine." That hasn't stopped it from soaring to the top spot on Apple's US App Store.