❌

Normal view

Received — 26 March 2026 ⏭

Not All Tokens Are Created Equal: Query-Efficient Jailbreak Fuzzing for LLMs

✇cs.AI, q-bio.NC updates on arXiv.org
By: Wenyu Chen · Xiangtao Meng · Chuanchao Zang · Li Wang · Xinyu Gao · Jianing Wang · Peng Zhan · Zheng Li · Shanqing Guo
25 March 2026 at 12:00
arXiv:2603.23269v1 Announce Type: cross Abstract: Large Language Models(LLMs) are widely deployed, yet are vulnerable to jailbreak prompts that elicit policy-violating outputs. Although prior studies have uncovered these risks, they typically treat all tokens as equally important during prompt mutation, overlooking the varying contributions of individual tokens to triggering model refusals. Consequently, these attacks introduce substantial redundant searching under query-constrained scenarios, reducing attack efficiency and hindering comprehensive vulnerability assessment. In this work, we conduct a token-level analysis of refusal behavior and observe that token contributions are highly skewed rather than uniform. Moreover, we find strong cross-model consistency in refusal tendencies, enabling the use of a surrogate model to estimate token-level contributions to the target model's refusals. Motivated by these findings, we propose TriageFuzz, a token-aware jailbreak fuzzing framework that adapts the fuzz testing approach with a series of customized designs. TriageFuzz leverages a surrogate model to estimate the contribution of individual tokens to refusal behaviors, enabling the identification of sensitive regions within the prompt. Furthermore, it incorporates a refusal-guided evolutionary strategy that adaptively weights candidate prompts with a lightweight scorer to steer the evolution toward bypassing safety constraints. Extensive experiments on six open-source LLMs and three commercial APIs demonstrate that TriageFuzz achieves comparable attack success rates (ASR) with significantly reduced query costs. Notably, it attains a 90% ASR with over 70% fewer queries compared to baselines. Even under an extremely restrictive budget of 25 queries, TriageFuzz outperforms existing methods, improving ASR by 20-40%.

SortedRL: Accelerating RL Training for LLMs through Online Length-Aware Scheduling

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yiqi Zhang · Huiqiang Jiang · Xufang Luo · Zhihe Yang · Chengruidong Zhang · Yifei Shen · Dongsheng Li · Yuqing Yang · Lili Qiu · Yang You
25 March 2026 at 12:00
arXiv:2603.23414v1 Announce Type: cross Abstract: Scaling reinforcement learning (RL) has shown strong promise for enhancing the reasoning abilities of large language models (LLMs), particularly in tasks requiring long chain-of-thought generation. However, RL training efficiency is often bottlenecked by the rollout phase, which can account for up to 70% of total training time when generating long trajectories (e.g., 16k tokens), due to slow autoregressive generation and synchronization overhead between rollout and policy updates. We propose SortedRL, an online length-aware scheduling strategy designed to address this bottleneck by improving rollout efficiency and maintaining training stability. SortedRL reorders rollout samples based on output lengths, prioritizing short samples forming groups for early updates. This enables large rollout batches, flexible update batches, and near on-policy micro-curriculum construction simultaneously. To further accelerate the pipeline, SortedRL incorporates a mechanism to control the degree of off-policy training through a cache-based mechanism, and is supported by a dedicated RL infrastructure that manages rollout and update via a stateful controller and rollout buffer. Experiments using LLaMA-3.1-8B and Qwen-2.5-32B on diverse tasks, including logical puzzles, and math challenges like AIME 24, Math 500, and Minerval, show that SortedRL reduces RL training bubble ratios by over 50%, while attaining 3.9% to 18.4% superior performance over baseline given same amount of data.

Cerebra: A Multidisciplinary AI Board for Multimodal Dementia Characterization and Risk Assessment

✇cs.AI, q-bio.NC updates on arXiv.org
By: Sheng Liu · Long Chen · Zeyun Zhao · Qinglin Gou · Qingyue Wei · Arjun Masurkar · Kevin M. Spiegler · Philip Kuball · Stefania C. Bray · Megan Bernath · Deanna R. Willis · Jiang Bian · Lei Xing · Eric Topol · Kyunghyun Cho · Yu Huang · Ruogu Fang · Narges Razavian · James Zou
25 March 2026 at 12:00
arXiv:2603.21597v2 Announce Type: replace Abstract: Modern clinical practice increasingly depends on reasoning over heterogeneous, evolving, and incomplete patient data. Although recent advances in multimodal foundation models have improved performance on various clinical tasks, most existing models remain static, opaque, and poorly aligned with real-world clinical workflows. We present Cerebra, an interactive multi-agent AI team that coordinates specialized agents for EHR, clinical notes, and medical imaging analysis. These outputs are synthesized into a clinician-facing dashboard that combines visual analytics with a conversational interface, enabling clinicians to interrogate predictions and contextualize risk at the point of care. Cerebra supports privacy-preserving deployment by operating on structured representations and remains robust when modalities are incomplete. We evaluated Cerebra using a massive multi-institutional dataset spanning 3 million patients from four independent healthcare systems. Cerebra consistently outperformed both state-of-the-art single-modality models and large multimodal language model baselines. In dementia risk prediction, it achieved AUROCs up to 0.80, compared with 0.74 for the strongest single-modality model and 0.68 for language model baselines. For dementia diagnosis, it achieved an AUROC of 0.86, and for survival prediction, a C-index of 0.81. In a reader study with experienced physicians, Cerebra significantly improved expert performance, increasing accuracy by 17.5 percentage points in prospective dementia risk estimation. These results demonstrate Cerebra's potential for interpretable, robust decision support in clinical care.

Dataset Distillation-based Hybrid Federated Learning on Non-IID Data

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiufang Shi · Wei Zhang · Yuheng Li · Mincheng Wu · Zhenyu Wen · Shibo He · Tejal Shah · Rajiv Ranjan
25 March 2026 at 12:00
arXiv:2409.17517v3 Announce Type: replace-cross Abstract: In federated learning, the heterogeneity of client data has a great impact on the performance of model training. Many heterogeneity issues in this process are raised by non-independently and identically distributed (non-IID) data. To address the issue of label distribution skew, we propose a hybrid federated learning framework called HFLDD, which integrates dataset distillation to generate approximately independent and equally distributed (IID) data, thereby improving the performance of model training. In particular, we partition the clients into heterogeneous clusters, where the data labels among different clients within a cluster are unbalanced while the data labels among different clusters are balanced. The cluster heads collect distilled data from the corresponding cluster members, and conduct model training in collaboration with the server. This training process is like traditional federated learning on IID data, and hence effectively alleviates the impact of non-IID data on model training. We perform a comprehensive analysis of the convergence behavior, communication overhead, and computational complexity of the proposed HFLDD. Extensive experimental results based on multiple public datasets demonstrate that when data labels are severely imbalanced, the proposed HFLDD outperforms the baseline methods in terms of both test accuracy and communication cost.

Graph Structure Learning with Privacy Guarantees for Open Graph Data

✇cs.AI, q-bio.NC updates on arXiv.org
By: Muhao Guo · Jiaqi Wu · Yizheng Liao · Wenke Lee · Shengzhe Chen · Yang Weng
25 March 2026 at 12:00
arXiv:2507.19116v3 Announce Type: replace-cross Abstract: Publishing open graph data while preserving individual privacy remains challenging when data publishers and data users are distinct entities. Although differential privacy (DP) provides rigorous guarantees, most existing approaches enforce privacy during model training rather than at the data publishing stage. This limits the applicability to open-data scenarios. We propose a privacy-preserving graph structure learning framework that integrates Gaussian Differential Privacy (GDP) directly into the data release process. Our mechanism injects structured Gaussian noise into raw data prior to publication and provides formal $\mu$-GDP guarantees, leading to tight $(\varepsilon, \delta)$-differential privacy bounds. Despite the distortion introduced by privatization, we prove that the original sparse inverse covariance structure can be recovered through an unbiased penalized likelihood formulation. We further extend the framework to discrete data using discrete Gaussian noise while preserving privacy guarantees. Extensive experiments on synthetic and real-world datasets demonstrate strong privacy-utility trade-offs, maintaining high graph recovery accuracy under rigorous privacy budgets. Our results establish a formal connection between differential privacy theory and privacy-preserving data publishing for graphical models.

Do Vision-Language Models Measure Up? Benchmarking Visual Measurement Reading with MeasureBench

✇cs.AI, q-bio.NC updates on arXiv.org
By: Fenfen Lin · Yesheng Liu · Haiyu Xu · Chen Yue · Zheqi He · Mingxuan Zhao · Miguel Hu Chen · Jiakang Liu · JG Yao · Xi Yang
25 March 2026 at 12:00
arXiv:2510.26865v2 Announce Type: replace-cross Abstract: Reading measurement instruments is effortless for humans and requires relatively little domain expertise, yet it remains surprisingly challenging for current vision-language models (VLMs) as we find in preliminary evaluation. In this work, we introduce MeasureBench, a benchmark on visual measurement reading covering both real-world and synthesized images of various types of measurements, along with an extensible pipeline for data synthesis. Our pipeline procedurally generates a specified type of gauge with controllable visual appearance, enabling scalable variation in key details such as pointers, scales, fonts, lighting, and clutter. Evaluation on popular proprietary and open-weight VLMs shows that even the strongest frontier VLMs struggle with measurement reading in general. We have also conducted preliminary experiments with reinforcement finetuning (RFT) over synthetic data, and find a significant improvement on both in-domain synthetic subset and real-world images. Our analysis highlights a fundamental limitation of current VLMs in fine-grained spatial grounding. We hope this resource and our code releases can help future advances on visually grounded numeracy and precise spatial perception of VLMs, bridging the gap between recognizing numbers and measuring the world.

Think Before You Drive: World Model-Inspired Multimodal Grounding for Autonomous Vehicles

✇cs.AI, q-bio.NC updates on arXiv.org
By: Haicheng Liao · Huanming Shen · Bonan Wang · Yongkang Li · Yihong Tang · Chengyue Wang · Dingyi Zhuang · Kehua Chen · Hai Yang · Chengzhong Xu · Zhenning Li
25 March 2026 at 12:00
arXiv:2512.03454v3 Announce Type: replace-cross Abstract: Interpreting natural-language commands to localize target objects is critical for autonomous driving (AD). Existing visual grounding (VG) methods for autonomous vehicles (AVs) typically struggle with ambiguous, context-dependent instructions, as they lack reasoning over 3D spatial relations and anticipated scene evolution. Grounded in the principles of world models, we propose ThinkDeeper, a framework that reasons about future spatial states before making grounding decisions. At its core is a Spatial-Aware World Model (SA-WM) that learns to reason ahead by distilling the current scene into a command-aware latent state and rolling out a sequence of future latent states, providing forward-looking cues for disambiguation. Complementing this, a hypergraph-guided decoder then hierarchically fuses these states with the multimodal input, capturing higher-order spatial dependencies for robust localization. In addition, we present DrivePilot, a multi-source VG dataset in AD, featuring semantic annotations generated by a Retrieval-Augmented Generation (RAG) and Chain-of-Thought (CoT)-prompted LLM pipeline. Extensive evaluations on six benchmarks, ThinkDeeper ranks #1 on the Talk2Car leaderboard and surpasses state-of-the-art baselines on DrivePilot, MoCAD, and RefCOCO/+/g benchmarks. Notably, it shows strong robustness and efficiency in challenging scenes (long-text, multi-agent, ambiguity) and retains superior performance even when trained on 50% of the data.

Spatial Omics in Gastrointestinal Oncology: Recent Advances, Therapeutic Insights, and Clinical Translation

✇(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Lin Zhong · Qian Li · Ting Xiong · Shengzheng Lin · Kaining Wang · Guanying Li · Ye Yan · Jianhao Liu · Chuansong Xue
23 March 2026 at 18:00

J Cancer. 2026 Jan 30;17(3):515-523. doi: 10.7150/jca.127381. eCollection 2026.

ABSTRACT

Gastrointestinal (GI) cancers remain a leading cause of cancer-related morbidity and mortality worldwide, largely due to their molecular heterogeneity, complex tumor microenvironment (TME), and variable treatment responses. In recent years, the emergence of spatially resolved omics technologies-encompassing spatial transcriptomics, proteomics, metabolomics, and epigenomics-has revolutionized the ability to interrogate tumor architecture with unprecedented resolution. These methods enable precise mapping of cellular and molecular interactions within intact tissue contexts, thereby uncovering spatially defined niches that influence tumor progression, immune evasion, and therapeutic resistance. In GI malignancies such as colorectal, gastric, and esophageal cancers, spatial omics have provided critical insights into cancer-stromal-immune crosstalk, identified predictive biomarkers for immunotherapy and targeted agents, and guided the development of novel therapeutic strategies. This review synthesizes the latest advances in spatial omics applied to GI oncology over the past five years, with an emphasis on their integration into early diagnosis, treatment stratification, and real-time monitoring of therapeutic efficacy. We also discuss current challenges, including standardization, data integration, and clinical validation, as well as future directions for incorporating spatial profiling into routine oncology practice. By bridging the gap between bench discoveries and bedside applications, spatial omics hold transformative potential for achieving truly personalized treatment in gastrointestinal cancers.

PMID:41869445 | PMC:PMC13003551 | DOI:10.7150/jca.127381

Integrated Machine Learning and Multi-Omics Identifies a Novel Molecular Signature for Improving the Prognosis of Hepatocellular Carcinoma

✇Omics in Hepatocellular
By: Zhongxu Wu · Jianguang Xiong · Qisheng Liu · Chengdang Wang · Dan Li · Liuliu Wei · Jian Ding
18 March 2026 at 18:00

J Hepatocell Carcinoma. 2026 Mar 11;13:574690. doi: 10.2147/JHC.S574690. eCollection 2026.

ABSTRACT

BACKGROUND: Hepatocellular carcinoma (HCC) exhibits significant molecular heterogeneity and complex immune microenvironment, which to some extent limits the accuracy of prognosis assessment and the formulation of individualized treatment strategies. This study aims to identify immune-derived molecular signatures based on multi-omics data and machine learning methods for the prognosis prediction and risk stratification of HCC.

METHODS: Based on weighted gene co-expression network analysis(WGCNA) and differential gene analysis,immune-derived molecular signature (IDMS) were screened in both single-cell and bulk transcriptomes. Prognostic model was constructed by multi-machine learning approachs. Subsequently, we investigated the differences in mutations, biological functions, and immune cell infiltration within the tumor microenvironment between the high- and low-risk groups.In addition, we comprehensively analyzed the drug sensitivity of IDMS and predicted potential drugs.

RESULTS: We identified seven hub genes at the single-cell and bulk transcriptome levels. Based on multiple machine learning, we constructed a prognostic model that demonstrated excellent performance in predicting overall survival for patients with HCC. IDMS -integrated normograms provide a promising and quantitative tool for clinical risk management.Notably, a significant difference in microsatellite instability (MSI) was observed between the high- and low-risk groups. This indicates that patients in the high-risk group might have a better response to immunotherapy. Additionally, we predicted potential drugs targeting to these risk subgroups.

CONCLUSION: Our research developed an IDMS that could serve as an effective tool for patient stratification management and prognosis prediction. This signature could provide a reference for immunotherapy for patients with HCC and improve their prognosis.

PMID:41847219 | PMC:PMC12991065 | DOI:10.2147/JHC.S574690

There are no more articles
❌