Large Language Models (LLMs) are typically trained to reflect a relatively uniform set of values, which limits their applicability to tasks that require understanding of nuanced human perspectives. Recent research has underscored the importance of enabling LLMs to support steerable pluralism -- the capacity to adopt a specific perspective and align generated outputs with it. In this work, we investigate whether Chain-of-Thought (CoT) reasoning techniques can be applied to building steerable pluralistic models. We explore several methods, including CoT prompting, fine-tuning on human-authored CoT, fine-tuning on synthetic explanations, and Reinforcement Learning with Verifiable Rewards (RLVR). We evaluate these approaches using the Value Kaleidoscope and OpinionQA datasets. Among the methods studied, RLVR consistently outperforms others and demonstrates strong training sample efficiency. We further analyze the generated CoT traces with respect to faithfulness and safety.

Post-training quantization for reducing the storage of deep neural network models has been demonstrated to be an effective way in various tasks. However, low-bit quantization while maintaining model accuracy is a challenging problem. In this paper, we present a range estimation method to improve the quantization performance for post-training quantization. We model the range estimation into an optimization problem of minimizing quantization errors by layer-wise local minima. We prove this problem is locally convex and present an efficient search algorithm to find the optimal solution. We propose the application of the above search algorithm to the transformed weights space to do further improvement in practice. Our experiments demonstrate that our method outperforms state-of-the-art performance generally on top-1 accuracy for image classification tasks on the ResNet series models and Inception-v3 model. The experimental results show that the proposed method has almost no loss of top-1 accuracy in 8-bit and 6-bit settings for image classifications, and the accuracy of 4-bit quantization is also significantly improved. The code is available at https://github.com/codeiscommitting/REQuant.

Large language models (LLMs) increasingly rely on Chain-of-Thought (CoT) prompting to improve problem-solving and provide seemingly transparent explanations. However, growing evidence shows that CoT often fail to faithfully represent the underlying reasoning process, raising concerns about their reliability in high-risk applications. Although prior studies have focused on mechanism-level analyses showing that CoTs can be unfaithful, they leave open the practical challenge of deciding whether a specific trajectory is faithful to the internal reasoning of the model. To address this gap, we introduce FaithCoT-Bench, a unified benchmark for instance-level CoT unfaithfulness detection. Our framework establishes a rigorous task formulation that formulates unfaithfulness detection as a discriminative decision problem, and provides FINE-CoT (Faithfulness instance evaluation for Chain-of-Thought), an expert-annotated collection of over 1,000 trajectories generated by four representative LLMs across four domains, including more than 300 unfaithful instances with fine-grained causes and step-level evidence. We further conduct a systematic evaluation of eleven representative detection methods spanning counterfactual, logit-based, and LLM-as-judge paradigms, deriving empirical insights that clarify the strengths and weaknesses of existing approaches and reveal the increased challenges of detection in knowledge-intensive domains and with more advanced models. To the best of our knowledge, FaithCoT-Bench establishes the first comprehensive benchmark for instance-level CoT faithfulness, setting a solid basis for future research toward more interpretable and trustworthy reasoning in LLMs.

Recent advances in image editing have shifted from manual pixel manipulation to employing deep learning methods like stable diffusion models, which now leverage cross-attention mechanisms for text-driven control. This transition has simplified the editing process but also introduced variability in results, such as inconsistent hair color changes. Our research aims to enhance the precision and reliability of prompt-to-prompt image editing frameworks by exploring and optimizing hyperparameters. We present a comprehensive study of the "word swap" method, develop an "attention re-weight method" for better adaptability, and propose the "CL P2P" framework to address existing limitations like cycle inconsistency. This work contributes to understanding and improving the interaction between hyperparameter settings and the architectural choices of neural network models, specifically their attention mechanisms, which significantly influence the composition and quality of the generated images.

Modern computer systems often rely on syslog, a simple, universal protocol that records every critical event across heterogeneous infrastructure. However, healthcare's rapidly growing clinical AI stack has no equivalent. As hospitals rush to pilot large language models and other AI-based clinical decision support tools, we still lack a standard way to record how, when, by whom, and for whom these AI models are used. Without that transparency and visibility, it is challenging to measure real-world performance and outcomes, detect adverse events, or correct bias or dataset drift. In the spirit of syslog, we introduce MedLog, a protocol for event-level logging of clinical AI. Any time an AI model is invoked to interact with a human, interface with another algorithm, or act independently, a MedLog record is created. This record consists of nine core fields: header, model, user, target, inputs, artifacts, outputs, outcomes, and feedback, providing a structured and consistent record of model activity. To encourage early adoption, especially in low-resource settings, and minimize the data footprint, MedLog supports risk-based sampling, lifecycle-aware retention policies, and write-behind caching; detailed traces for complex, agentic, or multi-stage workflows can also be captured under MedLog. MedLog can catalyze the development of new databases and software to store and analyze MedLog records. Realizing this vision would enable continuous surveillance, auditing, and iterative improvement of medical AI, laying the foundation for a new form of digital epidemiology.

Large language models (LLMs) have become increasingly popular in medical domains to assist physicians with a variety of clinical and operational tasks. Given the fast-paced and high-stakes environment of emergency departments (EDs), small language models (SLMs), characterized by a reduction in parameter count compared to LLMs, offer significant potential due to their inherent reasoning capability and efficient performance. This enables SLMs to support physicians by providing timely and accurate information synthesis, thereby improving clinical decision-making and workflow efficiency. In this paper, we present a comprehensive benchmark designed to identify SLMs suited for ED decision support, taking into account both specialized medical expertise and broad general problem-solving capabilities. In our evaluations, we focus on SLMs that have been trained on a mixture of general-domain and medical corpora. A key motivation for emphasizing SLMs is the practical hardware limitations, operational cost constraints, and privacy concerns in the typical real-world deployments. Our benchmark datasets include MedMCQA, MedQA-4Options, and PubMedQA, with the medical abstracts dataset emulating tasks aligned with real ED physicians' daily tasks. Experimental results reveal that general-domain SLMs surprisingly outperform their medically fine-tuned counterparts across these diverse benchmarks for ED. This indicates that for ED, specialized medical fine-tuning of the model may not be required.

The ongoing debate on whether reinforcement learning with verifiable rewards (RLVR) expands or shrinks the reasoning capabilities of large language models (LLMs) remains unresolved. Some studies contend that RLVR mainly improves sampling efficiency but at the expense of diversity and exploratory capacity, resulting in capability boundary shrinkage. In contrast, others demonstrate that prolonged training can lead to the emergence of novel reasoning strategies, suggesting capability boundary expansion. To reconcile these contradictory findings, we theoretically and empirically show that both perspectives are partially valid-each aligning with a separate phase in an inherent two-stage probability mass dynamic: (1) Exploitation stage: initially, the model primarily samples explored high-reward and low-reward tokens, while rarely selecting the potentially optimal token. Positive advantage estimates increase the probability of high-reward tokens and decrease those of low-reward tokens, yet the optimal token's probability remains largely unchanged during this stage. (2) Exploration stage: as training advances, the growth rate of previously acquired high-reward tokens slows as their probabilities approach saturation. When a potentially optimal token-now receiving positive advantage estimates-is occasionally sampled, its probability increases, while those of the originally high-reward tokens decrease. This dynamic suggests that over-exploitation during the exploitation stage may lead to capability boundary shrinkage, whereas prolonged training into the exploration stage can promote an expansion of the reasoning capability boundary. Building upon our insights, we revisit the potential of only using relative negative gradients for prolonging training, providing a theoretical and empirical foundation for the development of more advanced reasoning capabilities.

Recent advances in large language models (LLMs) have enabled a new class of AI agents that automate multiple stages of the data science workflow by integrating planning, tool use, and multimodal reasoning across text, code, tables, and visuals. This survey presents the first comprehensive, lifecycle-aligned taxonomy of data science agents, systematically analyzing and mapping forty-five systems onto the six stages of the end-to-end data science process: business understanding and data acquisition, exploratory analysis and visualization, feature engineering, model building and selection, interpretation and explanation, and deployment and monitoring. In addition to lifecycle coverage, we annotate each agent along five cross-cutting design dimensions: reasoning and planning style, modality integration, tool orchestration depth, learning and alignment methods, and trust, safety, and governance mechanisms. Beyond classification, we provide a critical synthesis of agent capabilities, highlight strengths and limitations at each stage, and review emerging benchmarks and evaluation practices. Our analysis identifies three key trends: most systems emphasize exploratory analysis, visualization, and modeling while neglecting business understanding, deployment, and monitoring; multimodal reasoning and tool orchestration remain unresolved challenges; and over 90% lack explicit trust and safety mechanisms. We conclude by outlining open challenges in alignment stability, explainability, governance, and robust evaluation frameworks, and propose future research directions to guide the development of robust, trustworthy, low-latency, transparent, and broadly accessible data science agents.

Diffusion large language models (dLLMs) are a new paradigm of non-autoregressive language models that are trained to predict multiple tokens in parallel and generate text via iterative unmasking. Recent works have successfully pretrained dLLMs to parity with autoregressive LLMs at the 8B scale, but dLLMs have yet to benefit from modern post-training techniques, e.g. reinforcement learning (RL), that have proven effective for autoregressive models. Crucially, algorithms designed for traditional LLMs aren't directly compatible with diffusion frameworks due to inherent differences in modeling assumptions. Moreover, existing attempts at dLLM post-training with RL rely on heuristic-based objectives with no theoretical grounding. In this work, we present Amortized Group Relative Policy Optimization (AGRPO), a principled on-policy RL algorithm designed specifically for dLLMs. AGRPO uses Monte Carlo sampling to compute an unbiased policy gradient estimate, making it the first tractable, faithful adaptation of policy gradient methods for dLLMs. We demonstrate AGRPO's effectiveness on different math/reasoning tasks, a common setting for RL with LLMs, achieving up to +7.6% absolute gain on GSM8K and 3.8x performance on the Countdown task over the baseline LLaDA-8B-Instruct model and 1.3x performance gains over comparable RL methods such as diffu-GRPO. Furthermore, these gains persist across different numbers of sampling steps at inference time, achieving better tradeoffs between compute and performance. Our results demonstrate that online RL algorithms can be extended to diffusion LLMs in principled ways, maintaining both theoretical soundness and practical effectiveness.

Foundation models for weather science are pre-trained on vast amounts of structured numerical data and outperform traditional weather forecasting systems. However, these models lack language-based reasoning capabilities, limiting their utility in interactive scientific workflows. Large language models (LLMs) excel at understanding and generating text but cannot reason about high-dimensional meteorological datasets. We bridge this gap by building a novel agentic framework for weather science. Our framework includes a Python code-based environment for agents (ZephyrusWorld) to interact with weather data, featuring tools like an interface to WeatherBench 2 dataset, geoquerying for geographical masks from natural language, weather forecasting, and climate simulation capabilities. We design Zephyrus, a multi-turn LLM-based weather agent that iteratively analyzes weather datasets, observes results, and refines its approach through conversational feedback loops. We accompany the agent with a new benchmark, ZephyrusBench, with a scalable data generation pipeline that constructs diverse question-answer pairs across weather-related tasks, from basic lookups to advanced forecasting, extreme event detection, and counterfactual reasoning. Experiments on this benchmark demonstrate the strong performance of Zephyrus agents over text-only baselines, outperforming them by up to 35 percentage points in correctness. However, on harder tasks, Zephyrus performs similarly to text-only baselines, highlighting the challenging nature of our benchmark and suggesting promising directions for future work.