arXiv:2605.24396v1 Announce Type: new Abstract: Long chains of thought (CoT) from current language models frequently contain logical gaps and unjustified leaps, limiting the gains from additional test-time compute. Improving reasoning quality directly would require process reward models, but the step-level annotations needed to train them are expensive and scarce. We find such a signal in how the model's confidence evolves during reasoning: premature confidence, the tendency to commit to an answer early and use the remaining tokens to rationalize it, strongly predicts flawed reasoning across tasks and model scales. We exploit this in progressive confidence shaping, a reinforcement learning objective that trains models to update their confidence as they reason rather than commit early -- rewarding gradual confidence growth and penalizing early commitment, with no external labels or reward models. The method improves accuracy and reasoning quality from 1.5B to 8B parameters across arithmetic (Countdown), math (DAPO, AIME), and science (ScienceQA): on Countdown, accuracy improves 3.2x (+42.0pp) and flawed reasoning drops 48pp; on AIME, Pass@64 improves 6.6pp. Consistent with this mechanism, the method also improves faithfulness: on a safety benchmark, our models more transparently surface misleading content in their reasoning traces rather than concealing it. Controlled experiments reveal that the problem and its remedy scale together: premature confidence grows with model size and task difficulty, and so do the gains from addressing it.

arXiv:2605.24171v1 Announce Type: cross Abstract: Large language models are increasingly used for vulnerability detection, yet their reliability under different prompt formulations remains uncharacterized. We present PromptAudit, a controlled evaluation framework that isolates prompt effects by fixing the dataset, decoding, and parsing while varying only the prompting strategy. Using five prompting strategies across five open-weight models on 1,000 CVEs (6,074 code samples spanning 16 programming languages), we evaluate accuracy, recall, abstention, coverage, and effective F1. We find that standard chain-of-thought prompting achieves the strongest overall operational performance, while few-shot prompting provides model-dependent benefits that are most pronounced for prompt-sensitive models. In contrast, adaptive chain-of-thought frequently suppresses recall and self-consistency induces excessive abstention, sharply reducing effective performance. These results show that vulnerability detection behavior is jointly determined by the model and the prompt, and that prompt sensitivity is a first-class system property that must be explicitly characterized in evaluation and deployment.

arXiv:2605.24384v1 Announce Type: cross Abstract: Language models (LMs) can exhibit systematic biases against speakers based on variations in their dialects, even in the absence of a dialect label, a behavior known as covert dialect bias. In this work, we quantify covert dialect bias in online discourse by evaluating how LMs associate stereotypical traits (derived from social psychology research on racial bias) with intent-equivalent tweets in Standard American English (SAE) and African-American Vernacular English (AAVE). While prior work shows that LMs associate more negative stereotypes with AAVE when evaluating tweets in isolation, we are surprised to find that this bias is significantly exacerbated when SAE / AAVE tweet pairs are compared side by side, a setting that more closely reflects high-impact decision making contexts in which models are used to rank candidates. The bias only worsens when dialect labels are explicitly specified. This is striking, given the extensive efforts from commercial developers to mitigate bias in their LMs. Encouragingly, we show that counterfactual fairness finetuning can mitigate covert dialect bias for some stereotypical traits, reducing average disparities when evaluating tweets in isolation, however, these improvements do not consistently hold across traits when evaluating SAE / AAVE tweets side by side. Our findings show that existing evaluation settings for covert dialect bias may underestimate its severity, specifically in contrastive settings. Additionally, overt dialect bias remains pronounced even after safety aligned finetuning, indicating that it remains an unresolved problem, and motivates the need for more robust evaluation and mitigation frameworks.

arXiv:2605.24831v1 Announce Type: cross Abstract: Non-Maximum Suppression (NMS) remains a key post-processing step in many real-time object detection pipelines, but it can introduce latency variation and deployment complexity in resource-constrained settings. Recent NMS-free designs such as YOLO26 aim to reduce this dependence through end-to-end detection, yet their performance relative to established NMS-based models such as YOLOv8 remains underexplored beyond standard benchmarks. This paper compares YOLOv8 and YOLO26 on Pascal VOC and VisDrone, representing general object detection and dense aerial small-object detection, respectively. Both model families are evaluated across five scales using accuracy, localization, model size, GFLOPs, and CPU/GPU latency. Results show that YOLO26 achieves stronger detection performance and lower model complexity on Pascal VOC across most scales, while the performance gap narrows on VisDrone, where both models struggle with dense small targets. YOLOv8 remains competitive in GPU latency, showing that NMS-free design does not guarantee universal deployment superiority. Overall, the study shows that detector selection depends on dataset characteristics, object scale, model capacity, and hardware constraints.

arXiv:2605.25198v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a powerful paradigm for improving language models on reasoning-intensive tasks, but its effectiveness is often limited by exploration. For example, models often fail on hard problems, leaving little useful reward signal. External expert traces offer a natural source of guidance, yet they may also expose reward-relevant content along the critical path to the verifier target, such as final answers, intermediate values, executable implementations, or answer-related entities. This content can create an unintended reward hacking channel, allowing the policy to obtain reward by copying the trace rather than learning the underlying reasoning or agentic behavior. Existing guided-RL methods reduce this risk by using partial trajectories, but they mainly control how much expert information is shown heuristically rather than which parts should be hidden. To this end, we propose Semantic Masked Expert Policy Optimization (SMEPO), a fine-grained semantic masking strategy for expert-guided RLVR. Instead of truncating traces coarsely or revealing them unchanged, SMEPO masks reward-relevant semantic spans along the critical path while preserving the expert's decomposition, plan, and procedural structure. This turns hard problems from reasoning from scratch into a fill-in-the-blank process: the policy can follow the expert's problem-solving route, but must still reconstruct the missing values, code, or entities by itself. SMEPO is simple to apply and requires no changes to the reward function or RL objective. Across diverse domains, including math, code, and agentic search, SMEPO improves accuracy by up to 3.2 points over GRPO and reduces training time by up to 4.2x. The code is available at https://github.com/mit-han-lab/SMEPO.

arXiv:2605.25252v1 Announce Type: cross Abstract: Reinforcement learning with verifiable rewards (RLVR) has become a standard paradigm for post-training language models, but in practice, verifiers are rarely perfect. Recent theoretical work predicts that verifier noise affects the rate of learning but not its final outcome, implying that sufficient compute should close any gap induced by imperfect supervision. We test this prediction empirically by post-training Qwen2.5 (0.5B, 1.5B) with GRPO on GSM8K while injecting controlled false-positive and false-negative noise into the binary correctness signal, and varying rollouts per prompt as a compute axis. In practice, the gap in validation accuracy persists under substantial compute scaling, with returns to compute that are sharply diminishing. We further find a structural asymmetry where false negatives monotonically degrade performance quicker than with false positives. These findings suggest verifier quality and training compute are not interchangeable, and that reducing false negatives is a more effective lever than scaling compute alone.

arXiv:2605.25254v1 Announce Type: cross Abstract: With unified model-generated images now widespread online, attributing their model of origin offers a path toward transparency and deeper insight into the characteristic behaviors of individual models. Prior work has explored provenance in LLM-generated text, diffusion model images, and datasets, but the separability of unified model-generated images remains an underexplored area. We address this gap by examining separability across corruption, domains, and prompt languages using images generated by seven unified models. We show that model attribution is highly feasible as our model achieves near-perfect accuracy with around 20K images per model. Corruptions and structural perturbations have only a modest effect on attribution performance, and cross-domain generalization reveals that semantic content contributes to separability but is not the dominant signal. Finally, we observe that for most models, prompt language attribution is around chance levels, suggesting minimal language-specific visual signatures. These findings highlight consistent model-specific visual characteristics in unified models outputs and open new directions for tracing and auditing generative image pipelines.

arXiv:2605.25278v1 Announce Type: cross Abstract: Understanding the statistics of level crossings in stochastic processes is crucial across many scientific disciplines. The traditional Kac-Rice formula gives the mean rate of level crossings and has found broad use. However, that mean rate captures only a coarse summary of the crossing process. It depends entirely on local properties of the stochastic process at a given instant and is therefore blind to the correlation structure of the process over time. To understand whether crossing events, such as neuronal spikes, tend to cluster in time, spread apart, or exhibit more complex temporal organization, one must go beyond the mean rate and study higher-order crossing statistics. Here we go beyond the mean by deriving the exact analytical formulae for the variance and Fano factor of arbitrary level crossings in smooth stationary Gaussian processes. Our exact solution reveals how the full temporal correlation structure dictates whether crossings cluster or become regular. In systems with oscillatory correlations, such as a stochastic damped harmonic oscillator, a recent crossing suppresses an immediate subsequent one, producing sub-Poissonian statistics. However, as damping increases and oscillations disappear, a large and slow excursion above the threshold can produce multiple closely spaced crossings, yielding super-Poissonian statistics. In purely relaxational, non-oscillatory systems, such as a mean-reverting process driven by Ornstein-Uhlenbeck noise, the competition between the timescales of the driving noise and system relaxation produces a richer landscape, including reentrant transitions between sub- and super-Poissonian statistics as the threshold level is varied. Taken together, the exact variance and Fano factor derived here complement the Kac-Rice mean rate, enabling more robust parameter estimation and model selection across any setting where Gaussian processes are used.

arXiv:2605.25440v1 Announce Type: cross Abstract: Verbal feedback delivered by attending surgeons in the operating room plays a critical formative role in resident trainee skill acquisition. Yet, assessing the quality of trainer feedback and its effectiveness in influencing trainee behavior during live surgery remains a challenge. Prior studies assessed feedback content relying on extensive manual annotation by expert human raters and focused on developing broad taxonomies that overlook the qualitative aspects of feedback delivery such as clarity or urgency. Limited existing automated methods, including keyword analysis and topic modeling, also fail to capture these nuanced aspects. We introduce a two-stage LLM-based framework that discovers interpretable feedback quality criteria grounded in the context of surgical training. Our method uses multi-agent prompting and surgical domain knowledge injection to discover a small set of human interpretable scoring criteria (e.g., Encouraging, Urgent, Clear). These criteria are then used to automatically score live surgical feedback via an LLM-as-a-judge approach. Evaluation on 4.2k trainer feedback instances demonstrates that our AI-discovered criteria outperform prior content-based frameworks in predicting feedback effectiveness, including observed trainee behavioral adjustments and trainer approval. This work advances scalable, human-aligned assessment of communication quality in the operating room and provides a foundation for improving surgical teaching practices.

arXiv:2511.12378v2 Announce Type: replace Abstract: Autonomous agents operating in sequential decision-making tasks under uncertainty can benefit from external action suggestions, which provide valuable guidance but inherently vary in reliability. Existing methods for incorporating such advice typically assume static and known suggester quality parameters, limiting practical deployment. We introduce a framework that dynamically learns and adapts to varying suggester reliability in partially observable environments. First, we integrate suggester quality directly into the agent's belief representation, enabling agents to infer and adjust their reliance on suggestions through Bayesian inference over suggester types. Second, we introduce an explicit ``ask'' action allowing agents to strategically request suggestions at critical moments, balancing informational gains against acquisition costs. Experimental evaluation demonstrates robust performance across varying suggester qualities, adaptation to changing reliability, and strategic management of suggestion requests. This work provides a foundation for adaptive human-agent collaboration by addressing suggestion uncertainty in uncertain environments.

arXiv:2511.20236v3 Announce Type: replace Abstract: Counterfactual explanations improve the actionable interpretability of machine learning models by identifying minimal changes required to achieve a desired outcome. However, existing methods often neglect dependencies among features, which can lead to unrealistic or impractical modifications. This limitation reduces the usefulness of counterfactual explanations in real-world decision-support systems. Motivated by applications in cybersecurity for email marketing, we propose DANCE (Diverse, Actionable, and Knowledge-Constrained Explanations), a method for generating counterfactuals that incorporate feature dependencies and domain constraints. DANCE models relationships between features using linear and probabilistic structures that can be learned from data or specified by experts. These dependencies are enforced during the search process to improve plausibility and feasibility. The method jointly optimizes plausibility, diversity, proximity, and sparsity within a unified objective. We evaluate DANCE on 140 datasets from OpenML and demonstrate that it achieves competitive or superior performance compared to existing approaches across multiple evaluation criteria. Additionally, we validate the method in a real-world industrial setting in collaboration with an email marketing platform, showing that it produces domain-consistent and actionable recommendations.

arXiv:2604.10783v2 Announce Type: replace Abstract: Designing reward functions for reinforcement learning (RL) in healthcare remains challenging because clinically meaningful outcomes are sparse, delayed, and difficult to explicitly specify. Although structured clinical data capture physiologic states, they often fail to reflect broader aspects of patient trajectories such as treatment response, recovery dynamics, and intervention burden. Clinical narratives, by contrast, encode longitudinal clinician assessments of disease progression, treatment effectiveness, and recovery, providing a potential source of trajectory-level supervision beyond predefined outcome metrics. We propose Clinical Narrative-informed Preference Rewards (CN-PR), a framework that learns reward functions directly from discharge summaries by treating clinical narratives as scalable supervision for trajectory-level preferences. Using a large language model, we derive trajectory quality scores and construct pairwise preferences between patient trajectories to learn rewards through preference-based optimization. To account for variability in narrative informativeness, we incorporate a task relevance signal that weights supervision according to its relevance to the downstream decision-making task. We evaluate CN-PR in dynamic sepsis treatment using offline RL. The learned reward demonstrated strong monotonic alignment with trajectory quality scores and produced policies associated with improved recovery-related outcomes, including increased organ support-free days and faster shock resolution, while maintaining mortality performance comparable to outcome-based reward baselines. These findings were preserved under external validation. Our results suggest that clinical narratives provide a scalable and expressive source of supervision for reward learning in dynamic treatment regimes.

arXiv:2605.11359v2 Announce Type: replace Abstract: Scientific data processing often requires task-specific algorithms or AI models, creating a barrier for domain scientists who need to analyze their data but may not have extensive computing or image-processing expertise. This barrier is especially pronounced when data are noisy, have a high dynamic range, are sparsely labeled, or are only loosely specified. We introduce CVEvolve, an autonomous agentic harness with a zero-code interface for scientific data-processing algorithm discovery. CVEvolve combines a multi-round search strategy with tools for code execution, evaluation implementation, history management, holdout testing, and optional inspection of scientific data and visual outputs. The search alternates between discovery and improvement actions, and uses lineage-aware stochastic candidate sampling to balance exploration and exploitation. We demonstrate CVEvolve on x-ray fluorescence microscopy image registration, Bragg peak detection, and high-energy diffraction microscopy image segmentation. Across these tasks, CVEvolve discovers algorithms that improve over baseline methods, while holdout test tracking helps identify candidates that generalize better than later over-optimized alternatives. These results show that zero-code, autonomous LLM-powered algorithm development can help domain scientists turn unstructured scientific image data into practical algorithms and downstream scientific discoveries.

arXiv:2602.07666v3 Announce Type: replace-cross Abstract: DARPA's AI Cyber Challenge (AIxCC, 2023--2025) is the largest competition to date for building fully autonomous cyber reasoning systems (CRSs) that leverage recent advances in AI -- particularly large language models (LLMs) -- to discover and remediate vulnerabilities in real-world open-source software. This paper presents the first systematic analysis of AIxCC. Drawing on design documents, source code, execution traces, and discussions with organizers and competing teams, we examine the competition's structure and key design decisions, characterize the architectural approaches of finalist CRSs, and analyze competition results beyond the final scoreboard. Our analysis reveals the factors that truly drove CRS performance, identifies genuine technical advances achieved by teams, and exposes limitations that remain open for future research. We conclude with lessons for organizing future competitions and broader insights toward deploying autonomous CRSs in practice.

Nature Medicine, Published online: 26 May 2026; doi:10.1038/s41591-026-04401-x

Biobanks are introducing trusted research environments to enable secure data access, yet they can increase costs and complexity for biomedical researchers who use them. We discuss challenges and solutions and introduce the Python Phenofhy community tool for Our Future Health.

Nature Medicine, Published online: 26 May 2026; doi:10.1038/s41591-026-04376-9

In a multicenter phase 2 trial, the fibroblast growth factor receptor inhibitor rogaratinib showed encouraging clinical efficacy in patients with succinate dehydrogenase-deficient gastrointestinal stromal tumors, suggesting a potential new treatment option for this patient population and demonstrating that an epigenetic mechanism of oncogene activation can be successfully targeted with a tyrosine kinase inhibitor.

Cell Death Discovery, Published online: 26 May 2026; doi:10.1038/s41420-026-03151-6

Reprogramming temozolomide response in glioblastoma through regulated and immunogenic cell death modalities

Nature Cancer, Published online: 25 May 2026; doi:10.1038/s43018-026-01165-8

In this phase 2 trial, Perez et al. reported the safety of the vitamin D receptor agonist paricalcitol with gemcitabine and nab-paclitaxel in patients with metastatic pancreatic ductal adenocarcinoma and the pharmacodynamic effects of paricalcitol in fibroblasts and immune cells.

Nature Medicine, Published online: 22 May 2026; doi:10.1038/s41591-026-04416-4

At week 68, in patients who experienced poor weight loss following bariatric surgery, semaglutide was associated with 18.0% weight loss compared to 0.4% weight gain in patients receiving placebo.

Nature Medicine, Published online: 22 May 2026; doi:10.1038/s41591-026-04438-y

Will medical trainees who rely on AI fail to develop foundational independent clinical reasoning? This Perspective outlines a precautionary framework to preserve foundational competence while supporting safe and effective AI integration in medical training.