The rapid advancement of multi-modal large reasoning models (MLRMs) -- enhanced versions of multimodal language models (MLLMs) equipped with reasoning capabilities -- has revolutionized diverse applications. However, their safety implications remain underexplored. While prior work has exposed critical vulnerabilities in unimodal reasoning models, MLRMs introduce distinct risks from cross-modal reasoning pathways. This work presents the first systematic safety analysis of MLRMs through large-scale empirical studies comparing MLRMs with their base MLLMs. Our experiments reveal three critical findings: (1) The Reasoning Tax: Acquiring reasoning capabilities catastrophically degrades inherited safety alignment. MLRMs exhibit 37.44% higher jailbreaking success rates than base MLLMs under adversarial attacks. (2) Safety Blind Spots: While safety degradation is pervasive, certain scenarios (e.g., Illegal Activity) suffer 25 times higher attack rates -- far exceeding the average 3.4 times increase, revealing scenario-specific vulnerabilities with alarming cross-model and datasets consistency. (3) Emergent Self-Correction: Despite tight reasoning-answer safety coupling, MLRMs demonstrate nascent self-correction -- 16.9% of jailbroken reasoning steps are overridden by safe answers, hinting at intrinsic safeguards. These findings underscore the urgency of scenario-aware safety auditing and mechanisms to amplify MLRMs' self-correction potential. To catalyze research, we open-source OpenSafeMLRM, the first toolkit for MLRM safety evaluation, providing unified interface for mainstream models, datasets, and jailbreaking methods. Our work calls for immediate efforts to harden reasoning-augmented AI, ensuring its transformative potential aligns with ethical safeguards.

As LLMs grow in capability, the task of supervising LLMs becomes more challenging. Supervision failures can occur if LLMs are sensitive to factors that supervisors are unaware of. We investigate Mechanistic Anomaly Detection (MAD) as a technique to augment supervision of capable models; we use internal model features to identify anomalous training signals so they can be investigated or discarded. We train detectors to flag points from the test environment that differ substantially from the training environment, and experiment with a large variety of detector features and scoring rules to detect anomalies in a set of "quirky" language models. We find that detectors can achieve high discrimination on some tasks, but no detector is effective across all models and tasks. MAD techniques may be effective in low-stakes applications, but advances in both detection and evaluation are likely needed if they are to be used in high stakes settings. Large language models (LLMs) excel at a wide range of tasks, yet as their capabilities grow, ensuring effective human oversight becomes increasingly challenging. These considerations have motivated Mechanistic Anomaly Detection (MAD): instead of attempting to understand all facets of a model's internal reasoning, MAD focuses on extracting a signal that indicates when an episode is anomalous--flagging behaviors that deviate from the expected patterns without assuming specific trigger mechanisms (Christiano, 2022). This signal could be used to filter out anomalous episodes, or to trigger deeper investigation.

Large Language Models (LLMs) have been applied to time series forecasting tasks, leveraging pre-trained language models as the backbone and incorporating textual data to purportedly enhance the comprehensive capabilities of LLMs for time series. However, are these texts really helpful for interpretation? This study seeks to investigate the actual efficacy and interpretability of such textual incorporations. Through a series of empirical experiments on textual prompts and textual prototypes, our findings reveal that the misalignment between two modalities exists, and the textual information does not significantly improve time series forecasting performance in many cases. Furthermore, visualization analysis indicates that the textual representations learned by existing frameworks lack sufficient interpretability when applied to time series data. We further propose a novel metric named Semantic Matching Index (SMI) to better evaluate the matching degree between time series and texts during our post hoc interpretability investigation. Our analysis reveals the misalignment and limited interpretability of texts in current time-series LLMs, and we hope this study can raise awareness of the interpretability of texts for time series. The code is available at https://github.com/zachysun/TS-Lang-Exp.

Enhancing the spatial perception capabilities of mobile robots is crucial for achieving embodied Vision-and-Language Navigation (VLN). Although significant progress has been made in simulated environments, directly transferring these capabilities to real-world scenarios often results in severe hallucination phenomena, causing robots to lose effective spatial awareness. To address this issue, we propose BrainNav, a bio-inspired spatial cognitive navigation framework inspired by biological spatial cognition theories and cognitive map theory. BrainNav integrates dual-map (coordinate map and topological map) and dual-orientation (relative orientation and absolute orientation) strategies, enabling real-time navigation through dynamic scene capture and path planning. Its five core modules-Hippocampal Memory Hub, Visual Cortex Perception Engine, Parietal Spatial Constructor, Prefrontal Decision Center, and Cerebellar Motion Execution Unit-mimic biological cognitive functions to reduce spatial hallucinations and enhance adaptability. Validated in a zero-shot real-world lab environment using the Limo Pro robot, BrainNav, compatible with GPT-4, outperforms existing State-of-the-Art (SOTA) Vision-and-Language Navigation in Continuous Environments (VLN-CE) methods without fine-tuning.

Estimating item difficulty through field-testing is often resource-intensive and time-consuming. As such, there is strong motivation to develop methods that can predict item difficulty at scale using only the item content. Large Language Models (LLMs) represent a new frontier for this goal. The present research examines the feasibility of using an LLM to predict item difficulty for K-5 mathematics and reading assessment items (N = 5170). Two estimation approaches were implemented: (a) a direct estimation method that prompted the LLM to assign a single difficulty rating to each item, and (b) a feature-based strategy where the LLM extracted multiple cognitive and linguistic features, which were then used in ensemble tree-based models (random forests and gradient boosting) to predict difficulty. Overall, direct LLM estimates showed moderate to strong correlations with true item difficulties. However, their accuracy varied by grade level, often performing worse for early grades. In contrast, the feature-based method yielded stronger predictive accuracy, with correlations as high as r = 0.87 and lower error estimates compared to both direct LLM predictions and baseline regressors. These findings highlight the promise of LLMs in streamlining item development and reducing reliance on extensive field testing and underscore the importance of structured feature extraction. We provide a seven-step workflow for testing professionals who would want to implement a similar item difficulty estimation approach with their item pool.

Named Entity Recognition (NER), a fundamental task in Natural Language Processing (NLP), has shown significant advancements for high-resource languages. However, due to a lack of annotated datasets and limited representation in Pre-trained Language Models (PLMs), it remains understudied and challenging for low-resource languages. To address these challenges, we propose a data augmentation technique that generates culturally plausible sentences and experiments on four low-resource Pakistani languages; Urdu, Shahmukhi, Sindhi, and Pashto. By fine-tuning multilingual masked Large Language Models (LLMs), our approach demonstrates significant improvements in NER performance for Shahmukhi and Pashto. We further explore the capability of generative LLMs for NER and data augmentation using few-shot learning.

The recent advancements in Large Language Models (LLMs) have generated considerable interest in their utilization for sequential recommendation tasks. While collaborative signals from similar users are central to recommendation modeling, effectively transforming these signals into a format that LLMs can understand and utilize remains challenging. The critical challenges include selecting relevant demonstrations from large-scale user interactions and ensuring their alignment with LLMs' reasoning process. To address these challenges, we introduce AdaptRec, a self-adaptive fram-ework that leverages LLMs for sequential recommendations by incorporating explicit collaborative signals. AdaptRec employs a two-phase user selection mechanism -- User Similarity Retrieval and Self-Adaptive User Selection -- to efficiently identify relevant user sequences in large-scale datasets from multi-metric evaluation. We also develop a User-Based Similarity Retrieval Prompt, enabling the model to actively select similar users and continuously refine its selection criteria during training. Using the collaborative signals from similar users, we construct a User-Contextualized Recommendation Prompt that translates their behavior sequences into natural language, explicitly integrating this information into the recommendation process. Experiments demonstrate AdaptRec's superior performance, with significant improvements in HitRatio@1 scores of 7.13\%, 18.16\%, and 10.41\% across real-world datasets with full fine-tuning, and even higher gains of 23.00\%, 15.97\%, and 17.98\% in few-shot scenarios.

This paper introduces SLOs-Serve, a system designed for serving multi-stage large language model (LLM) requests with application- and stage-specific service level objectives (SLOs). The key idea behind SLOs-Serve is to customize the allocation of tokens to meet these SLO requirements. SLOs-Serve uses a multi-SLO dynamic programming-based algorithm to continuously optimize token allocations under SLO constraints by exploring the full design space of chunked prefill and (optional) speculative decoding. Leveraging this resource planning algorithm, SLOs-Serve effectively supports multi-SLOs and multi-replica serving with dynamic request routing while being resilient to bursty arrivals. Our evaluation across 6 LLM application scenarios (including summarization, coding, chatbot, tool calling, and reasoning) demonstrates that SLOs-Serve improves per-GPU serving capacity by 2.2x on average compared to prior state-of-the-art systems.

With the development of Large Language Models (LLMs), numerous benchmarks have been proposed to measure and compare the capabilities of different LLMs. However, evaluating LLMs is costly due to the large number of test instances and their slow inference speed. In this paper, we aim to explore how to efficiently estimate a model's real performance on a given benchmark based on its evaluation results on a small number of instances sampled from the benchmark. Inspired by Collaborative Filtering (CF) in Recommendation Systems (RS), we treat LLMs as users and test instances as items and propose a two-stage method. In the first stage, we treat instance selection as recommending products to users to choose instances that can easily distinguish model performance. In the second stage, we see performance prediction as rating prediction problem in RS to predict the target LLM's behavior on unselected instances. Experiments on multiple LLMs and datasets imply that our method can accurately estimate the target model's performance while largely reducing its inference overhead.

ABSTRACT The growing complexity of construction management (CM) projects, coupled with challenges such as strict regulatory requirements and labor shortages, requires specialized analytical tools that streamline project workflow and enhance performance. Although large language models (LLMs) have demonstrated exceptional performance in general reasoning tasks, their effectiveness in tackling CM-specific challenges, such as precise quantitative analysis and regulatory interpretation, remains inadequately explored. To bridge this gap, this study introduces CMExamSet, a comprehensive benchmarking dataset comprising 689 authentic multiple-choice questions sourced from 1 arXiv:2504.08779v1 The results indicate that GPT-4o and Claude 3.7 surpass typical human pass thresholds (70%), with average accuracies of 82% and 83%, respectively. Additionally, both models performed better on single-step tasks, with accuracies of 85.7% (GPT-4o) and 86.7% (Claude 3.7). Multi-step tasks were more challenging, reducing performance to 76.5% and 77.6%, respectively. Our error pattern analysis further reveals that conceptual misunderstandings are the most common (44.4% and 47.9%), underscoring the need for enhanced domain-specific reasoning models. These findings underscore the potential of LLMs as valuable supplementary analytical tools in CM, while highlighting the need for domain-specific refinements and sustained human oversight in complex decision making. INTRODUCTION The construction industry is undergoing a transformation driven by digital technologies, increased project complexity, heterogeneous regulations, and ongoing labor shortages (Abioye et al. 2021). These changes create a pressing need for intelligent tools that can augment human expertise and support decision-making in construction management (CM) (Regona et al. 2022). Among these technologies, large language models (LLMs) such as GPT-4 and Claude have shown a comparative performance in general reasoning, natural language understanding, and educational applications (Ooi et al. 2025).