Autonomous driving systems remain brittle in rare, ambiguous, and out-of-distribution scenarios, where human driver succeed through contextual reasoning. Shared autonomy has emerged as a promising approach to mitigate such failures by incorporating human input when autonomy is uncertain. However, most existing methods restrict arbitration to low-level trajectories, which represent only geometric paths and therefore fail to preserve the underlying driving intent. We propose a unified shared autonomy framework that integrates human input and autonomous planners at a higher level of abstraction. Our method leverages Vision Language Models (VLMs) to infer driver intent from multi-modal cues -- such as driver actions and environmental context -- and to synthesize coherent strategies that mediate between human and autonomous control. We first study the framework in a mock-human setting, where it achieves perfect recall alongside high accuracy and precision. A human-subject survey further shows strong alignment, with participants agreeing with arbitration outcomes in 92% of cases. Finally, evaluation on the Bench2Drive benchmark demonstrates a substantial reduction in collision rate and improvement in overall performance compared to pure autonomy. Arbitration at the level of semantic, language-based representations emerges as a design principle for shared autonomy, enabling systems to exercise common-sense reasoning and maintain continuity with human intent.

LLMs can perform multi-step reasoning through Chain-of-Thought (CoT), but they cannot reliably verify their own logic. Even when they reach correct answers, the underlying reasoning may be flawed, undermining trust in high-stakes scenarios. To mitigate this issue, we introduce VeriCoT, a neuro-symbolic method that extracts and verifies formal logical arguments from CoT reasoning. VeriCoT formalizes each CoT reasoning step into first-order logic and identifies premises that ground the argument in source context, commonsense knowledge, or prior reasoning steps. The symbolic representation enables automated solvers to verify logical validity while the NL premises allow humans and systems to identify ungrounded or fallacious reasoning steps. Experiments on the ProofWriter, LegalBench, and BioASQ datasets show VeriCoT effectively identifies flawed reasoning, and serves as a strong predictor of final answer correctness. We also leverage VeriCoT's verification signal for (1) inference-time self-reflection, (2) supervised fine-tuning (SFT) on VeriCoT-distilled datasets and (3) preference fine-tuning (PFT) with direct preference optimization (DPO) using verification-based pairwise rewards, further improving reasoning validity and accuracy.

Chain-of-Thought (CoT) prompting is a key technique for enabling complex reasoning in large language models. However, generating full, fixed-length rationales is computationally wasteful, inflating both token usage and latency. We introduce LEASH: Logit-Entropy Adaptive Stopping Heuristic, a training-free decoding algorithm that adaptively halts rationale generation. LEASH monitors two intrinsic signals: the slope of token-level entropy and the improvement in the top-logit margin. It terminates the generation once both signals plateau, indicating the model has reached a stable reasoning state. Across four instruction-tuned models on the GSM8K and AQuA-RAT benchmarks, LEASH reduces average token generation by 30--35% and latency by 27%, while incurring a 10 p.p. accuracy drop relative to CoT. LEASH is model-agnostic and requires no additional training or supervision, offering a simple and efficient alternative to CoT decoding.

A central feature of many deliberative processes, such as citizens' assemblies and deliberative polls, is the opportunity for participants to engage directly with experts. While participants are typically invited to propose questions for expert panels, only a limited number can be selected due to time constraints. This raises the challenge of how to choose a small set of questions that best represent the interests of all participants. We introduce an auditing framework for measuring the level of representation provided by a slate of questions, based on the social choice concept known as justified representation (JR). We present the first algorithms for auditing JR in the general utility setting, with our most efficient algorithm achieving a runtime of $O(mn\log n)$, where $n$ is the number of participants and $m$ is the number of proposed questions. We apply our auditing methods to historical deliberations, comparing the representativeness of (a) the actual questions posed to the expert panel (chosen by a moderator), (b) participants' questions chosen via integer linear programming, (c) summary questions generated by large language models (LLMs). Our results highlight both the promise and current limitations of LLMs in supporting deliberative processes. By integrating our methods into an online deliberation platform that has been used for over hundreds of deliberations across more than 50 countries, we make it easy for practitioners to audit and improve representation in future deliberations.

1. A hard stop for the implementation of rigorous conservation initiatives is our lack of key species data, especially occurrence data. Furthermore, researchers have to contend with an accelerated speed at which new information must be collected and processed due to anthropogenic activity. Publications ranging from scientific papers to gray literature contain this crucial information but their data are often not machine-readable, requiring extensive human work to be retrieved. 2. We present the ARETE R package, an open-source software aiming to automate data extraction of species occurrences powered by large language models, namely using the chatGPT Application Programming Interface. This R package integrates all steps of the data extraction and validation process, from Optical Character Recognition to detection of outliers and output in tabular format. Furthermore, we validate ARETE through systematic comparison between what is modelled and the work of human annotators. 3. We demonstrate the usefulness of the approach by comparing range maps produced using GBIF data and with those automatically extracted for 100 species of spiders. Newly extracted data allowed to expand the known Extent of Occurrence by a mean three orders of magnitude, revealing new areas where the species were found in the past, which mayhave important implications for spatial conservation planning and extinction risk assessments. 4. ARETE allows faster access to hitherto untapped occurrence data, a potential game changer in projects requiring such data. Researchers will be able to better prioritize resources, manually verifying selected species while maintaining automated extraction for the majority. This workflow also allows predicting available bibliographic data during project planning.

"Thinking with Text" and "Thinking with Images" paradigm significantly improve the reasoning ability of large language models (LLMs) and Vision Language Models (VLMs). However, these paradigms have inherent limitations. (1) Images capture only single moments and fail to represent dynamic processes or continuous changes, and (2) The separation of text and vision as distinct modalities, hindering unified multimodal understanding and generation. To overcome these limitations, we introduce "Thinking with Video", a new paradigm that leverages video generation models, such as Sora-2, to bridge visual and textual reasoning in a unified temporal framework. To support this exploration, we developed the Video Thinking Benchmark (VideoThinkBench). VideoThinkBench encompasses two task categories: (1) vision-centric tasks (e.g., Eyeballing Puzzles), and (2) text-centric tasks (e.g., subsets of GSM8K, MMMU). Our evaluation establishes Sora-2 as a capable reasoner. On vision-centric tasks, Sora-2 is generally comparable to state-of-the-art (SOTA) VLMs, and even surpasses VLMs on several tasks, such as Eyeballing Games. On text-centric tasks, Sora-2 achieves 92% accuracy on MATH, and 75.53% accuracy on MMMU. Furthermore, we systematically analyse the source of these abilities. We also find that self-consistency and in-context learning can improve Sora-2's performance. In summary, our findings demonstrate that the video generation model is the potential unified multimodal understanding and generation model, positions "thinking with video" as a unified multimodal reasoning paradigm.

Developing accurate biomedical Question Answering (QA) systems in low-resource languages remains a major challenge, limiting equitable access to reliable medical knowledge. This paper introduces BanglaMedQA and BanglaMMedBench, the first large-scale Bangla biomedical Multiple Choice Question (MCQ) datasets designed to evaluate reasoning and retrieval in medical artificial intelligence (AI). The study applies and benchmarks several Retrieval-Augmented Generation (RAG) strategies, including Traditional, Zero-Shot Fallback, Agentic, Iterative Feedback, and Aggregate RAG, combining textbook-based and web retrieval with generative reasoning to improve factual accuracy. A key novelty lies in integrating a Bangla medical textbook corpus through Optical Character Recognition (OCR) and implementing an Agentic RAG pipeline that dynamically selects between retrieval and reasoning strategies. Experimental results show that the Agentic RAG achieved the highest accuracy 89.54% with openai/gpt-oss-120b, outperforming other configurations and demonstrating superior rationale quality. These findings highlight the potential of RAG-based methods to enhance the reliability and accessibility of Bangla medical QA, establishing a foundation for future research in multilingual medical artificial intelligence.

Modeling user preferences across domains remains a key challenge in slate recommendation (i.e. recommending an ordered sequence of items) research. We investigate how Large Language Models (LLM) can effectively act as world models of user preferences through pairwise reasoning over slates. We conduct an empirical study involving several LLMs on three tasks spanning different datasets. Our results reveal relationships between task performance and properties of the preference function captured by LLMs, hinting towards areas for improvement and highlighting the potential of LLMs as world models in recommender systems.

As the role of Large Language Models (LLM)-based coding assistants in software development becomes more critical, so does the role of the bugs they generate in the overall cybersecurity landscape. While a number of LLM code security benchmarks have been proposed alongside approaches to improve the security of generated code, it remains unclear to what extent they have impacted widely used coding LLMs. Here, we show that even the latest open-weight models are vulnerable in the earliest reported vulnerability scenarios in a realistic use setting, suggesting that the safety-functionality trade-off has until now prevented effective patching of vulnerabilities. To help address this issue, we introduce a new severity metric that reflects the risk posed by an LLM-generated vulnerability, accounting for vulnerability severity, generation chance, and the formulation of the prompt that induces vulnerable code generation - Prompt Exposure (PE). To encourage the mitigation of the most serious and prevalent vulnerabilities, we use PE to define the Model Exposure (ME) score, which indicates the severity and prevalence of vulnerabilities a model generates.

When a language model generates text, the selection of individual tokens might lead it down very different reasoning paths, making uncertainty difficult to quantify. In this work, we consider whether reasoning language models represent the alternate paths that they could take during generation. To test this hypothesis, we use hidden activations to control and predict a language model's uncertainty during chain-of-thought reasoning. In our experiments, we find a clear correlation between how uncertain a model is at different tokens, and how easily the model can be steered by controlling its activations. This suggests that activation interventions are most effective when there are alternate paths available to the model -- in other words, when it has not yet committed to a particular final answer. We also find that hidden activations can predict a model's future outcome distribution, demonstrating that models implicitly represent the space of possible paths.