Large language models (LLMs) have evolved into agentic systems capable of autonomous tool use and multi-step reasoning for complex problem-solving. However, post-training approaches building upon general-purpose foundation models consistently underperform in agentic tasks, particularly in open-source implementations. We identify the root cause: the absence of robust agentic foundation models forces models during post-training to simultaneously learn diverse agentic behaviors while aligning them to expert demonstrations, thereby creating fundamental optimization tensions. To this end, we are the first to propose incorporating Agentic Continual Pre-training (Agentic CPT) into the deep research agents training pipeline to build powerful agentic foundational models. Based on this approach, we develop a deep research agent model named AgentFounder. We evaluate our AgentFounder-30B on 10 benchmarks and achieve state-of-the-art performance while retains strong tool-use ability, notably 39.9% on BrowseComp-en, 43.3% on BrowseComp-zh, and 31.5% Pass@1 on HLE.

Transcending human cognitive limitations represents a critical frontier in LLM training. Proprietary agentic systems like DeepResearch have demonstrated superhuman capabilities on extremely complex information-seeking benchmarks such as BrowseComp, a feat previously unattainable. We posit that their success hinges on a sophisticated reasoning pattern absent in open-source models: the ability to systematically reduce extreme uncertainty when navigating vast information landscapes. Based on this insight, we introduce WebSailor, a complete post-training methodology designed to instill this crucial capability. Our approach involves generating novel, high-uncertainty tasks through structured sampling and information obfuscation, RFT cold start, and an efficient agentic RL training algorithm, Duplicating Sampling Policy Optimization (DUPO). With this integrated pipeline, WebSailor significantly outperforms all open-source agents in complex information-seeking tasks, matching proprietary agents' performance and closing the capability gap.

Unraveling the dynamical motions of biomolecules is essential for bridging their structure and function, yet it remains a major computational challenge. Molecular dynamics (MD) simulation provides a detailed depiction of biomolecular motion, but its high-resolution temporal evolution comes at significant computational cost, limiting its applicability to timescales of biological relevance. Deep learning approaches have emerged as promising solutions to overcome these computational limitations by learning to predict long-timescale dynamics. However, generalizable kinetics models for proteins remain largely unexplored, and the fundamental limits of achievable acceleration while preserving dynamical accuracy are poorly understood. In this work, we fill this gap with DeepJump, an Euclidean-Equivariant Flow Matching-based model for predicting protein conformational dynamics across multiple temporal scales. We train DeepJump on trajectories of the diverse proteins of mdCATH, systematically studying our model's performance in generalizing to long-term dynamics of fast-folding proteins and characterizing the trade-off between computational acceleration and prediction accuracy. We demonstrate the application of DeepJump to ab initio folding, showcasing prediction of folding pathways and native states. Our results demonstrate that DeepJump achieves significant $\approx$1000$\times$ computational acceleration while effectively recovering long-timescale dynamics, providing a stepping stone for enabling routine simulation of proteins.

Charts are a crucial visual medium for communicating and representing information. While Large Vision-Language Models (LVLMs) have made progress on chart question answering (CQA), the task remains challenging, particularly when models attend to irrelevant regions of the chart. In this work, we present ChartGaze, a new eye-tracking dataset that captures human gaze patterns during chart reasoning tasks. Through a systematic comparison of human and model attention, we find that LVLMs often diverge from human gaze, leading to reduced interpretability and accuracy. To address this, we propose a gaze-guided attention refinement that aligns image-text attention with human fixations. Our approach improves both answer accuracy and attention alignment, yielding gains of up to 2.56 percentage points across multiple models. These results demonstrate the promise of incorporating human gaze to enhance both the reasoning quality and interpretability of chart-focused LVLMs.

BACKGROUND: Most artificial intelligence tools used to estimate nutritional content rely on image input. However, whether large language models (LLMs) can accurately predict nutritional values based solely on text descriptions of foods consumed remains unknown. If effective, this approach could enable simpler dietary monitoring without the need for photographs. METHODS: We used 24-hour dietary recalls from adolescents aged 12-19 years in the National Health and Nutrition Examination Survey (NHANES). An open-source quantized LLM was prompted using a 10-shot, chain-of-thought approach to estimate energy and five macronutrients based solely on text strings listing foods and their quantities. We then applied parameter-efficient fine-tuning (PEFT) to evaluate whether predictive accuracy improved. NHANES-calculated values served as the ground truth for energy, proteins, carbohydrates, total sugar, dietary fiber and total fat. RESULTS: In a pooled dataset of 11,281 adolescents (49.9% male, mean age 15.4 years), the vanilla LLM yielded poor predictions. The mean absolute error (MAE) was 652.08 for energy and the Lin's CCC <0.46 across endpoints. In contrast, the fine-tuned model performed substantially better, with energy MAEs ranging from 171.34 to 190.90 across subsets, and Lin's CCC exceeding 0.89 for all outcomes. CONCLUSIONS: When prompted using a chain-of-thought approach and fine-tuned with PEFT, open-source LLMs exposed solely to text input can accurately predict energy and macronutrient values from 24-hour dietary recalls. This approach holds promise for low-burden, text-based dietary monitoring tools.

Several video understanding tasks, such as natural language temporal video grounding, temporal activity localization, and audio description generation, require "temporally dense" reasoning over frames sampled at high temporal resolution. However, computing frame-level features for these tasks is computationally expensive given the temporal resolution requirements. In this paper, we make three contributions to reduce the cost of computing features for temporally dense tasks. First, we introduce a vision transformer (ViT) architecture, dubbed ResidualViT, that leverages the large temporal redundancy in videos to efficiently compute temporally dense frame-level features. Our architecture incorporates (i) learnable residual connections that ensure temporal consistency across consecutive frames and (ii) a token reduction module that enhances processing speed by selectively discarding temporally redundant information while reusing weights of a pretrained foundation model. Second, we propose a lightweight distillation strategy to approximate the frame-level features of the original foundation model. Finally, we evaluate our approach across four tasks and five datasets, in both zero-shot and fully supervised settings, demonstrating significant reductions in computational cost (up to 60%) and improvements in inference speed (up to 2.5x faster), all while closely approximating the accuracy of the original foundation model.

Large Language Models (LLMs) are increasingly deployed in roles requiring nuanced psychological understanding, such as emotional support agents, counselors, and decision-making assistants. However, their ability to interpret human personality traits, a critical aspect of such applications, remains unexplored, particularly in ecologically valid conversational settings. While prior work has simulated LLM "personas" using discrete Big Five labels on social media data, the alignment of LLMs with continuous, ground-truth personality assessments derived from natural interactions is largely unexamined. To address this gap, we introduce a novel benchmark comprising semi-structured interview transcripts paired with validated continuous Big Five trait scores. Using this dataset, we systematically evaluate LLM performance across three paradigms: (1) zero-shot and chain-of-thought prompting with GPT-4.1 Mini, (2) LoRA-based fine-tuning applied to both RoBERTa and Meta-LLaMA architectures, and (3) regression using static embeddings from pretrained BERT and OpenAI's text-embedding-3-small. Our results reveal that all Pearson correlations between model predictions and ground-truth personality traits remain below 0.26, highlighting the limited alignment of current LLMs with validated psychological constructs. Chain-of-thought prompting offers minimal gains over zero-shot, suggesting that personality inference relies more on latent semantic representation than explicit reasoning. These findings underscore the challenges of aligning LLMs with complex human attributes and motivate future work on trait-specific prompting, context-aware modeling, and alignment-oriented fine-tuning.

Large language models (LLMs) now solve multi-step problems by emitting extended chains of thought. During the process, they often re-derive the same intermediate steps across problems, inflating token usage and latency. This saturation of the context window leaves less capacity for exploration. We study a simple mechanism that converts recurring reasoning fragments into concise, reusable "behaviors" (name + instruction) via the model's own metacognitive analysis of prior traces. These behaviors are stored in a "behavior handbook" which supplies them to the model in-context at inference or distills them into parameters via supervised fine-tuning. This approach achieves improved test-time reasoning across three different settings - 1) Behavior-conditioned inference: Providing the LLM relevant behaviors in-context during reasoning reduces number of reasoning tokens by up to 46% while matching or improving baseline accuracy; 2) Behavior-guided self-improvement: Without any parameter updates, the model improves its own future reasoning by leveraging behaviors from its own past problem solving attempts. This yields up to 10% higher accuracy than a naive critique-and-revise baseline; and 3) Behavior-conditioned SFT: SFT on behavior-conditioned reasoning traces is more effective at converting non-reasoning models into reasoning models as compared to vanilla SFT. Together, these results indicate that turning slow derivations into fast procedural hints enables LLMs to remember how to reason, not just what to conclude.

Diabetic retinopathy (DR) is a leading cause of blindness worldwide, and AI systems can expand access to fundus photography screening. Current FDA-cleared systems primarily provide binary referral outputs, where this minimal output may limit clinical trust and utility. Yet, determining the most effective output format to enhance clinician-AI performance is an empirical challenge that is difficult to assess at scale. We evaluated multimodal large language models (MLLMs) for DR detection and their ability to simulate clinical AI assistance across different output types. Two models were tested on IDRiD and Messidor-2: GPT-4o, a general-purpose MLLM, and MedGemma, an open-source medical model. Experiments included: (1) baseline evaluation, (2) simulated AI assistance with synthetic predictions, and (3) actual AI-to-AI collaboration where GPT-4o incorporated MedGemma outputs. MedGemma outperformed GPT-4o at baseline, achieving higher sensitivity and AUROC, while GPT-4o showed near-perfect specificity but low sensitivity. Both models adjusted predictions based on simulated AI inputs, but GPT-4o's performance collapsed with incorrect ones, whereas MedGemma remained more stable. In actual collaboration, GPT-4o achieved strong results when guided by MedGemma's descriptive outputs, even without direct image access (AUROC up to 0.96). These findings suggest MLLMs may improve DR screening pipelines and serve as scalable simulators for studying clinical AI assistance across varying output configurations. Open, lightweight models such as MedGemma may be especially valuable in low-resource settings, while descriptive outputs could enhance explainability and clinician trust in clinical workflows.

Traditional backdoor attacks in federated learning (FL) operate within constrained attack scenarios, as they depend on visible triggers and require physical modifications to the target object, which limits their practicality. To address this limitation, we introduce a novel backdoor attack prototype for FL called the out-of-distribution (OOD) backdoor attack ($\mathtt{OBA}$), which uses OOD data as both poisoned samples and triggers simultaneously. Our approach significantly broadens the scope of backdoor attack scenarios in FL. To improve the stealthiness of $\mathtt{OBA}$, we propose $\mathtt{SoDa}$, which regularizes both the magnitude and direction of malicious local models during local training, aligning them closely with their benign versions to evade detection. Empirical results demonstrate that $\mathtt{OBA}$ effectively circumvents state-of-the-art defenses while maintaining high accuracy on the main task. To address this security vulnerability in the FL system, we introduce $\mathtt{BNGuard}$, a new server-side defense method tailored against $\mathtt{SoDa}$. $\mathtt{BNGuard}$ leverages the observation that OOD data causes significant deviations in the running statistics of batch normalization layers. This allows $\mathtt{BNGuard}$ to identify malicious model updates and exclude them from aggregation, thereby enhancing the backdoor robustness of FL. Extensive experiments across various settings show the effectiveness of $\mathtt{BNGuard}$ on defending against $\mathtt{SoDa}$. The code is available at https://github.com/JiiahaoXU/SoDa-BNGuard.