❌

Normal view

Received — 26 March 2026 ⏭

Three Creates All: You Only Sample 3 Steps

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yuren Cai · Guangyi Wang · Zongqing Li · Li Li · Zhihui Liu · Songzhi Su
25 March 2026 at 12:00
arXiv:2603.22375v1 Announce Type: cross Abstract: Diffusion models deliver high-fidelity generation but remain slow at inference time due to many sequential network evaluations. We find that standard timestep conditioning becomes a key bottleneck for few-step sampling. Motivated by layer-dependent denoising dynamics, we propose Multi-layer Time Embedding Optimization (MTEO), which freeze the pretrained diffusion backbone and distill a small set of step-wise, layer-wise time embeddings from reference trajectories. MTEO is plug-and-play with existing ODE solvers, adds no inference-time overhead, and trains only a tiny fraction of parameters. Extensive experiments across diverse datasets and backbones show state-of-the-art performance in the few-step sampling and substantially narrow the gap between distillation-based and lightweight methods. Code will be available.

Genomic atlas of Bifidobacterium infantis and B. longum informs infant probiotic design

✇Cell
By: Yan Shao · Shuyi Wang · Bonface M. Gichuki · Mark D. Stares · Timothy J. Rozday · Nitin Kumar · Hilary P. Browne · Nicholas J.R. Dawson · James M. Njunge · Caroline Tigoi · Narshion Ngao · Mohammod Jobayer Chisti · Benson O. Singa · Samuel Kariuki · Abdoulaye Hama Diallo · Ali Faisal Saleem · Syed Asad Ali · Ezekiel Mupere · Emmie Mbale · Kirkby D. Tickell · Wieger P. Voskuijl · Christina L. Lancioni · Robert H.J. Bandsma · Tahmeed Ahmed · Judd L. Walson · James A. Berkley · Trevor D. Lawley
18 February 2026 at 08:00
A global genomic survey of infant gut bifidobacteria shows that B. infantis remains highly prevalent and diverse in infants from low- and middle-income countries but scarce in Western, industrialized populations and poorly represented in current probiotics. This genomic and culture collection catalogs geo-specific B. infantis strains and provides a blueprint for developing probiotics tailored to local diets and populations to support infant health.
Received — 17 March 2026 ⏭

SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiangyi Li · Wenbo Chen · Yimin Liu · Shenghan Zheng · Xiaokun Chen · Yifeng He · Yubo Li · Bingran You · Haotian Shen · Jiankai Sun · Shuyi Wang · Binxu Li · Qunhong Zeng · Di Wang · Xuandong Zhao · Yuanli Wang · Roey Ben Chaim · Zonglin Di · Yipeng Gao · Junwei He · Yizhuo He · Liqiang Jing · Luyang Kong · Xin Lan · Jiachen Li · Songlin Li · Yijiang Li · Yueqian Lin · Xinyi Liu · Xuanqing Liu · Haoran Lyu · Ze Ma · Bowei Wang · Runhui Wang · Tianyu Wang · Wengao Ye · Yue Zhang · Hanwen Xing · Yiqi Xue · Steven Dillmann · Han-chung Lee
16 March 2026 at 12:00
arXiv:2602.12670v3 Announce Type: replace Abstract: Agent Skills are structured packages of procedural knowledge that augment LLM agents at inference time. Despite rapid adoption, there is no standard way to measure whether they actually help. We present SkillsBench, a benchmark of 86 tasks across 11 domains paired with curated Skills and deterministic verifiers. Each task is evaluated under three conditions: no Skills, curated Skills, and self-generated Skills. We test 7 agent-model configurations over 7,308 trajectories. Curated Skills raise average pass rate by 16.2 percentage points(pp), but effects vary widely by domain (+4.5pp for Software Engineering to +51.9pp for Healthcare) and 16 of 84 tasks show negative deltas. Self-generated Skills provide no benefit on average, showing that models cannot reliably author the procedural knowledge they benefit from consuming. Focused Skills with 2--3 modules outperform comprehensive documentation, and smaller models with Skills can match larger models without them.
Received — 14 March 2026 ⏭

CircRNA-encoded RIPK1-98 protein drives lung adenocarcinoma progression

✇(Multiomics OR Omics) AND (Lung OR gastric OR Hepatocellular)
By: Qi Sun · Runqiu Chi · Xiao Zhang · Tianxiang Chen · Xin Xu · Lifang Ma · Xiaoting Tian · Yayou Miao · Yongchun Yu · Rui Kang · Guido Kroemer · Yi Shi · Daolin Tang · Jiayi Wang
13 March 2026 at 18:00

Dev Cell. 2026 Mar 12:S1534-5807(26)00079-1. doi: 10.1016/j.devcel.2026.02.014. Online ahead of print.

ABSTRACT

Unexplored biological matter-including uncharacterized genetic elements, molecular entities, and microbial components-remains poorly understood. Here, we use integrated multi-omics approaches to identify and characterize previously unrecognized protein products encoded by circular RNAs (circRNAs) in human tissue specimens and to delineate their roles in the progression of lung adenocarcinoma (LUAD). The transcription of precursor mRNA by RNA polymerase Ⅱ subunit A (RPB1) is crucial for the biogenesis of these potential circRNA-encoded proteins. Functional and translational analyses link their expression to distinct pathological stages of LUAD in patients. The protein RIPK1-98, encoded by circRIPK1, was identified as functionally distinct from its parental gene product, receptor-interacting serine/threonine kinase 1 (RIPK1). RIPK1-98 modulates cyclin-dependent kinase 2 (CDK2)-dependent cell-cycle regulation, thereby facilitating tumor proliferation in cellular and animal models. Together, these findings suggest that RIPK1-98 serves as a biomarker for cell-cycle progression in LUAD and highlight its potential as a therapeutic target to counteract resistance to first-line treatments, such as osimertinib.

PMID:41825439 | DOI:10.1016/j.devcel.2026.02.014

CircRNA-encoded RIPK1-98 protein drives lung adenocarcinoma progression

✇Omics In Lung
By: Qi Sun · Runqiu Chi · Xiao Zhang · Tianxiang Chen · Xin Xu · Lifang Ma · Xiaoting Tian · Yayou Miao · Yongchun Yu · Rui Kang · Guido Kroemer · Yi Shi · Daolin Tang · Jiayi Wang
13 March 2026 at 18:00

Dev Cell. 2026 Mar 12:S1534-5807(26)00079-1. doi: 10.1016/j.devcel.2026.02.014. Online ahead of print.

ABSTRACT

Unexplored biological matter-including uncharacterized genetic elements, molecular entities, and microbial components-remains poorly understood. Here, we use integrated multi-omics approaches to identify and characterize previously unrecognized protein products encoded by circular RNAs (circRNAs) in human tissue specimens and to delineate their roles in the progression of lung adenocarcinoma (LUAD). The transcription of precursor mRNA by RNA polymerase Ⅱ subunit A (RPB1) is crucial for the biogenesis of these potential circRNA-encoded proteins. Functional and translational analyses link their expression to distinct pathological stages of LUAD in patients. The protein RIPK1-98, encoded by circRIPK1, was identified as functionally distinct from its parental gene product, receptor-interacting serine/threonine kinase 1 (RIPK1). RIPK1-98 modulates cyclin-dependent kinase 2 (CDK2)-dependent cell-cycle regulation, thereby facilitating tumor proliferation in cellular and animal models. Together, these findings suggest that RIPK1-98 serves as a biomarker for cell-cycle progression in LUAD and highlight its potential as a therapeutic target to counteract resistance to first-line treatments, such as osimertinib.

PMID:41825439 | DOI:10.1016/j.devcel.2026.02.014

Received — 11 March 2026 ⏭

SkillsBench: Benchmarking How Well Agent Skills Work Across Diverse Tasks

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xiangyi Li · Wenbo Chen · Yimin Liu · Shenghan Zheng · Xiaokun Chen · Yifeng He · Yubo Li · Bingran You · Haotian Shen · Jiankai Sun · Shuyi Wang · Binxu Li · Qunhong Zeng · Di Wang · Xuandong Zhao · Yuanli Wang · Roey Ben Chaim · Zonglin Di · Yipeng Gao · Junwei He · Yizhuo He · Liqiang Jing · Luyang Kong · Xin Lan · Jiachen Li · Songlin Li · Yijiang Li · Yueqian Lin · Xinyi Liu · Xuanqing Liu · Haoran Lyu · Ze Ma · Bowei Wang · Runhui Wang · Tianyu Wang · Wengao Ye · Yue Zhang · Hanwen Xing · Yiqi Xue · Steven Dillmann · Han-chung Lee
10 March 2026 at 12:00
arXiv:2602.12670v2 Announce Type: replace Abstract: Agent Skills are structured packages of procedural knowledge that augment LLM agents at inference time. Despite rapid adoption, there is no standard way to measure whether they actually help. We present SkillsBench, a benchmark of 86 tasks across 11 domains paired with curated Skills and deterministic verifiers. Each task is evaluated under three conditions: no Skills, curated Skills, and self-generated Skills. We test 7 agent-model configurations over 7,308 trajectories. Curated Skills raise average pass rate by 16.2 percentage points(pp), but effects vary widely by domain (+4.5pp for Software Engineering to +51.9pp for Healthcare) and 16 of 84 tasks show negative deltas. Self-generated Skills provide no benefit on average, showing that models cannot reliably author the procedural knowledge they benefit from consuming. Focused Skills with 2--3 modules outperform comprehensive documentation, and smaller models with Skills can match larger models without them.
Received — 5 March 2026 ⏭

SycoEval-EM: Sycophancy Evaluation of Large Language Models in Simulated Clinical Encounters for Emergency Care

✇cs.AI, q-bio.NC updates on arXiv.org
By: Dongshen Peng · Yi Wang · Austin Schoeffler · Carl Preiksaitis · Christian Rose
5 March 2026 at 13:00
arXiv:2601.16529v2 Announce Type: replace Abstract: Large language models (LLMs) show promise in clinical decision support yet risk acquiescing to patient pressure for inappropriate care. We introduce SycoEval-EM, a multi-agent simulation framework evaluating LLM robustness through adversarial patient persuasion in emergency medicine. Across 20 LLMs and 1,875 encounters spanning three Choosing Wisely scenarios, acquiescence rates ranged from 0-100\%. Models showed higher vulnerability to imaging requests (38.8\%) than opioid prescriptions (25.0\%), with model capability poorly predicting robustness. All persuasion tactics proved equally effective (30.0-36.0\%), indicating general susceptibility rather than tactic-specific weakness. Our findings demonstrate that static benchmarks inadequately predict safety under social pressure, necessitating multi-turn adversarial testing for clinical AI certification.

Generalized Discrete Diffusion with Self-Correction

✇cs.AI, q-bio.NC updates on arXiv.org
By: Linxuan Wang · Ziyi Wang · Yikun Bai · Wei Deng · Guang Lin · Qifan Song
4 March 2026 at 13:00
arXiv:2603.02230v1 Announce Type: cross Abstract: Self-correction is an effective technique for maintaining parallel sampling in discrete diffusion models with minimal performance degradation. Prior work has explored self-correction at inference time or during post-training; however, such approaches often suffer from limited generalization and may impair reasoning performance. GIDD pioneers pretraining-based self-correction via a multi-step BERT-style uniform-absorbing objective. However, GIDD relies on a continuous interpolation-based pipeline with opaque interactions between uniform transitions and absorbing masks, which complicates hyperparameter tuning and hinders practical performance. In this work, we propose a Self-Correcting Discrete Diffusion (SCDD) model to reformulate pretrained self-correction with explicit state transitions and learn directly in discrete time. Our framework also simplifies the training noise schedule, eliminates a redundant remasking step, and relies exclusively on uniform transitions to learn self-correction. Experiments at the GPT-2 scale demonstrate that our method enables more efficient parallel decoding while preserving generation quality.

MASPOB: Bandit-Based Prompt Optimization for Multi-Agent Systems with Graph Neural Networks

✇cs.AI, q-bio.NC updates on arXiv.org
By: Zhi Hong · Qian Zhang · Jiahang Sun · Zhiwei Shang · Mingze Kong · Xiangyi Wang · Yao Shu · Zhongxiang Dai
4 March 2026 at 13:00
arXiv:2603.02630v1 Announce Type: cross Abstract: Large Language Models (LLMs) have achieved great success in many real-world applications, especially the one serving as the cognitive backbone of Multi-Agent Systems (MAS) to orchestrate complex workflows in practice. Since many deployment scenarios preclude MAS workflow modifications and its performance is highly sensitive to the input prompts, prompt optimization emerges as a more natural approach to improve its performance. However, real-world prompt optimization for MAS is impeded by three key challenges: (1) the need of sample efficiency due to prohibitive evaluation costs, (2) topology-induced coupling among prompts, and (3) the combinatorial explosion of the search space. To address these challenges, we introduce MASPOB (Multi-Agent System Prompt Optimization via Bandits), a novel sample-efficient framework based on bandits. By leveraging Upper Confidence Bound (UCB) to quantify uncertainty, the bandit framework balances exploration and exploitation, maximizing gains within a strictly limited budget. To handle topology-induced coupling, MASPOB integrates Graph Neural Networks (GNNs) to capture structural priors, learning topology-aware representations of prompt semantics. Furthermore, it employs coordinate ascent to decompose the optimization into univariate sub-problems, reducing search complexity from exponential to linear. Extensive experiments across diverse benchmarks demonstrate that MASPOB achieves state-of-the-art performance, consistently outperforming existing baselines.

Leverage Knowledge Graph and Large Language Model for Law Article Recommendation: A Case Study of Chinese Criminal Law

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yongming Chen · Miner Chen · Ye Zhu · Juan Pei · Siyu Chen · Yu Zhou · Yi Wang · Yifan Zhou · Hao Li · Songan Zhang
4 March 2026 at 13:00
arXiv:2410.04949v3 Announce Type: replace-cross Abstract: Judicial efficiency is critical to social stability. However, in many countries worldwide, grassroots courts face substantial case backlogs, and judicial decisions remain heavily dependent on judges' cognitive efforts, with insufficient intelligent tools to enhance efficiency. To address this issue, we propose a highly efficient law article recommendation approach combining a Knowledge Graph (KG) and a Large Language Model (LLM). First, we construct a Case-Enhanced Law Article Knowledge Graph (CLAKG) to store current law articles, historical case information, and their interconnections, alongside an LLM-based automated construction method. Building on this, we propose a closed-loop law article recommendation framework integrating graph embedding-based retrieval and KG-grounded LLM reasoning. Experiments on judgment documents from China Judgments Online demonstrate that our method boosts law article recommendation accuracy from 0.549 to 0.694, outperforming strong baselines significantly. To support reproducibility and future research, all source code and processed datasets are publicly available on GitHub (see Data Availability Statement).

LEDOM: Reverse Language Model

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xunjian Yin · Sitao Cheng · Yuxi Xie · Xinyu Hu · Li Lin · Xinyi Wang · Liangming Pan · William Yang Wang · Xiaojun Wan
4 March 2026 at 13:00
arXiv:2507.01335v3 Announce Type: replace-cross Abstract: Autoregressive language models are trained exclusively left-to-right. We explore the complementary factorization, training right-to-left at scale, and ask what reasoning patterns emerge when a model conditions on future context to predict the past. We train LEDOM, an open-source purely reverse autoregressive language model (2B/7B parameters, 435B tokens), and find it develops capabilities distinct from forward models, including abductive inference, question synthesis, and natural resolution of the reversal curse. We then explore one application of the reverse model: combining forward likelihood $P(y \mid x)$ with reverse posterior $P(x \mid y)$ through noisy channel duality. We propose Reverse Reward, which reranks forward outputs using reverse posterior estimates, and prove that bidirectional scoring penalizes hallucinated reasoning chains whose backward reconstruction degrades. Reverse Reward yields gains of up to 6.6\% on AIME 2024 and 15\% on AMC 2023 across multiple strong baselines. We release all models, code, and data here.

xLLM Technical Report

✇cs.AI, q-bio.NC updates on arXiv.org
By: Tongxuan Liu · Tao Peng · Peijun Yang · Xiaoyang Zhao · Xiusheng Lu · Weizhe Huang · Zirui Liu · Xiaoyu Chen · Zhiwei Liang · Jun Xiong · Donghe Jin · Minchao Zhang · Jinrong Guo · Yingxu Deng · Xu Zhang · Xianzhe Dong · Siqi Wang · Siyu Wu · Yu Wu · Zihan Tang · Yuting Zeng · Yanshu Wang · Jinguang Liu · Meng Kang · Menxin Li · Yunlong Wang · Yiming Liu · Xiaolong Ma · Yifan Wang · Yichen Zhang · Jinrun Yin · Keyang Zheng · Jiawei Yin · Jun Zhang · Ziyue Wang · Xiaobo Lin · Liangyu Liu · Liwei Lan · Yang Liu · Chunhua Peng · Han Liu · Songcheng Ren · Xuezhu Wang · Yunheng Shen · Yi Wang · Guyue Liu · Yitao Hu · Hui Chen · Tong Yang · Hailong Yang · Jing Li · Guiguang Ding · Ke Zhang
4 March 2026 at 13:00
arXiv:2510.14686v2 Announce Type: replace-cross Abstract: We introduce xLLM, an intelligent and efficient Large Language Model (LLM) inference framework designed for high-performance, large-scale enterprise-grade serving, with deep optimizations for diverse AI accelerators. To address these challenges, xLLM builds a novel decoupled service-engine architecture. At the service layer, xLLM-Service features an intelligent scheduling module that efficiently processes multimodal requests and co-locates online and offline tasks through unified elastic scheduling to maximize cluster utilization. This module also relies on a workload-adaptive dynamic Prefill-Decode (PD) disaggregation policy and a novel Encode-Prefill-Decode (EPD) disaggregation policy designed for multimodal inputs. Furthermore, it incorporates a distributed architecture to provide global KV Cache management and robust fault-tolerant capabilities for high availability. At the engine layer, xLLM-Engine co-optimizes system and algorithm designs to fully saturate computing resources. This is achieved through comprehensive multi-layer execution pipeline optimizations, an adaptive graph mode and an xTensor memory management. xLLM-Engine also further integrates algorithmic enhancements such as optimized speculative decoding and dynamic EPLB, collectively serving to substantially boost throughput and inference efficiency. Extensive evaluations demonstrate that xLLM delivers significantly superior performance and resource efficiency. Under identical TPOT constraints, xLLM achieves throughput up to 1.7x that of MindIE and 2.2x that of vLLM-Ascend with Qwen-series models, while maintaining an average throughput of 1.7x that of MindIE with Deepseek-series models. xLLM framework is publicly available at https://github.com/jd-opensource/xllm and https://github.com/jd-opensource/xllm-service.
Received — 25 February 2026 ⏭

AdaptStress: Online Adaptive Learning for Interpretable and Personalized Stress Prediction Using Multivariate and Sparse Physiological Signals

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xueyi Wang · Claudine J. C. Lamoth · Elisabeth Wilhelm
24 February 2026 at 13:00
arXiv:2602.18521v1 Announce Type: cross Abstract: Continuous stress forecasting could potentially contribute to lifestyle interventions. This paper presents a novel, explainable, and individualized approach for stress prediction using physiological data from consumer-grade smartwatches. We develop a time series forecasting model that leverages multivariate features, including heart rate variability, activity patterns, and sleep metrics, to predict stress levels across 16 temporal horizons (History window: 3, 5, 7, 9 days; forecasting window: 1, 3, 5, 7 days). Our evaluation involves 16 participants monitored for 10-15 weeks. We evaluate our approach across 16 participants, comparing against state-of-the-art time series models (Informer, TimesNet, PatchTST) and traditional baselines (CNN, LSTM, CNN-LSTM) across multiple temporal horizons. Our model achieved performance with an MSE of 0.053, MAE of 0.190, and RMSE of 0.226 in optimal settings (5-day input, 1-day prediction). A comparison with the baseline models shows that our model outperforms TimesNet, PatchTST, CNN-LSTM, LSTM, and CNN under all conditions, representing improvements of 36.9%, 25.5%, and 21.5% over the best baseline. According to the explanability analysis, sleep metrics are the most dominant and consistent stress predictors (importance: 1.1, consistency: 0.9-1.0), while activity features exhibit high inter-participant variability (0.1-0.2). Most notably, the model captures individual-specific patterns where identical features can have opposing effects across users, validating its personalization capabilities. These findings establish that consumer wearables, combined with adaptive and interpretable deep learning, can deliver relevant stress assessment adapted to individual physiological responses, providing a foundation for scalable, continuous, explainable mental health monitoring in real-world settings.

Rethinking the Design of Reinforcement Learning-Based Deep Research Agents

✇cs.AI, q-bio.NC updates on arXiv.org
By: Yi Wan · Jiuqi Wang · Liam Li · Jinsong Liu · Ruihao Zhu · Zheqing Zhu
24 February 2026 at 13:00
arXiv:2510.15862v4 Announce Type: replace Abstract: Large language models (LLMs) augmented with external tools are increasingly deployed as deep research agents that gather, reason over, and synthesize web information to answer complex queries. Although recent open-source systems achieve strong empirical performance via reinforcement learning from web interactions, the impact of key design choices remains under-explored. We formalize deep research as reinforcement learning in an episodic finite Markov decision process and construct a competitive baseline agent grounded in this formulation. Building on this foundation, we systematically examine critical design decisions at both training and inference time and identify four factors that substantially improve performance: replacing rule-based rewards with AI feedback from an LLM judge, fine-tuning with the on-policy RLOO algorithm instead of the off-policy GRPO algorithm, filtering low-quality training samples, and employing an error-tolerant test-time rollout strategy. Together, these design choices yield a deep research agent that establishes state-of-the-art performance among 7B-scale agents when evaluated across ten widely used benchmarks.

Adaptive Runge-Kutta Dynamics for Spatiotemporal Prediction

✇cs.AI, q-bio.NC updates on arXiv.org
By: Xuanle Zhao · Yue Sun · Ziyi Wang · Bo Xu · Tielin Zhang
24 February 2026 at 13:00
arXiv:2405.14504v2 Announce Type: replace-cross Abstract: Spatiotemporal prediction is important in solving natural problems and processing video frames, especially in weather forecasting and human action recognition. Recent advances attempt to incorporate prior physical knowledge into the deep learning framework to estimate the unknown governing partial differential equations (PDEs) in complex dynamics, which have shown promising results in spatiotemporal prediction tasks. However, previous approaches only restrict neural network architectures or loss functions to acquire physical or PDE features, which decreases the representative capacity of a neural network. Meanwhile, the updating process of the physical state cannot be effectively estimated. To solve the problems mentioned above, we introduce a physical-guided neural network, which utilizes an adaptive second-order Runge-Kutta method with physical constraints to model the physical states more precisely. Furthermore, we propose a frequency-enhanced Fourier module to strengthen the model's ability to estimate the spatiotemporal dynamics. We evaluate our model on both spatiotemporal and video prediction tasks. The experimental results show that our model outperforms several state-of-the-art methods and performs the best in several spatiotemporal scenarios with a much smaller parameter count.

RAIR: A Rule-Aware Benchmark Uniting Challenging Long-Tail and Visual Salience Subset for E-commerce Relevance Assessment

✇cs.AI, q-bio.NC updates on arXiv.org
By: Chenji Lu · Zhuo Chen · Hui Zhao · Zhenyi Wang · Pengjie Wang · Chuan Yu · Jian Xu
24 February 2026 at 13:00
arXiv:2512.24943v2 Announce Type: replace-cross Abstract: Search relevance plays a central role in web e-commerce. While large language models (LLMs) have shown significant results on relevance task, existing benchmarks lack sufficient complexity for comprehensive model assessment, resulting in an absence of standardized relevance evaluation metrics across the industry. To address this limitation, we propose Rule-Aware benchmark with Image for Relevance assessment(RAIR), a Chinese dataset derived from real-world scenarios. RAIR established a standardized framework for relevance assessment and provides a set of universal rules, which forms the foundation for standardized evaluation. Additionally, RAIR analyzes essential capabilities required for current relevance models and introduces a comprehensive dataset consists of three subset: (1) a general subset with industry-balanced sampling to evaluate fundamental model competencies; (2) a long-tail hard subset focus on challenging cases to assess performance limits; (3) a visual salience subset for evaluating multimodal understanding capabilities. We conducted experiments on RAIR using 14 open and closed-source models. The results demonstrate that RAIR presents sufficient challenges even for GPT-5, which achieved the best performance. RAIR data are now available, serving as an industry benchmark for relevance assessment while providing new insights into general LLM and Visual Language Model(VLM) evaluation.

Step 3.5 Flash: Open Frontier-Level Intelligence with 11B Active Parameters

✇cs.AI, q-bio.NC updates on arXiv.org
By: Ailin Huang · Ang Li · Aobo Kong · Bin Wang · Binxing Jiao · Bo Dong · Bojun Wang · Boyu Chen · Brian Li · Buyun Ma · Chang Su · Changxin Miao · Changyi Wan · Chao Lou · Chen Hu · Chen Xu · Chenfeng Yu · Chengting Feng · Chengyuan Yao · Chunrui Han · Dan Ma · Dapeng Shi · Daxin Jiang · Dehua Ma · Deshan Sun · Di Qi · Enle Liu · Fajie Zhang · Fanqi Wan · Guanzhe Huang · Gulin Yan · Guoliang Cao · Guopeng Li · Han Cheng · Hangyu Guo · Hanshan Zhang · Hao Nie · Haonan Jia · Haoran Lv · Hebin Zhou · Hekun Lv · Heng Wang · Heung-Yeung Shum · Hongbo Huang · Hongbo Peng · Hongyu Zhou · Hongyuan Wang · Houyong Chen · Huangxi Zhu · Huimin Wu · Huiyong Guo · Jia Wang · Jian Zhou · Jianjian Sun · Jiaoren Wu · Jiaran Zhang · Jiashu Lv · Jiashuo Liu · Jiayi Fu · Jiayu Liu · Jie Cheng · Jie Luo · Jie Yang · Jie Zhou · Jieyi Hou · Jing Bai · Jingcheng Hu · Jingjing Xie · Jingwei Wu · Jingyang Zhang · Jishi Zhou · Junfeng Liu · Junzhe Lin · Ka Man Lo · Kai Liang · Kaibo Liu · Kaijun Tan · Kaiwen Yan · Kaixiang Li · Kang An · Kangheng Lin · Lei Yang · Liang Lv · Liang Zhao · Liangyu Chen · Lieyu Shi · Liguo T
24 February 2026 at 13:00
arXiv:2602.10604v2 Announce Type: replace-cross Abstract: We introduce Step 3.5 Flash, a sparse Mixture-of-Experts (MoE) model that bridges frontier-level agentic intelligence and computational efficiency. We focus on what matters most when building agents: sharp reasoning and fast, reliable execution. Step 3.5 Flash pairs a 196B-parameter foundation with 11B active parameters for efficient inference. It is optimized with interleaved 3:1 sliding-window/full attention and Multi-Token Prediction (MTP-3) to reduce the latency and cost of multi-round agentic interactions. To reach frontier-level intelligence, we design a scalable reinforcement learning framework that combines verifiable signals with preference feedback, while remaining stable under large-scale off-policy training, enabling consistent self-improvement across mathematics, code, and tool use. Step 3.5 Flash demonstrates strong performance across agent, coding, and math tasks, achieving 85.4% on IMO-AnswerBench, 86.4% on LiveCodeBench-v6 (2024.08-2025.05), 88.2% on tau2-Bench, 69.0% on BrowseComp (with context management), and 51.0% on Terminal-Bench 2.0, comparable to frontier models such as GPT-5.2 xHigh and Gemini 3.0 Pro. By redefining the efficiency frontier, Step 3.5 Flash provides a high-density foundation for deploying sophisticated agents in real-world industrial environments.
Received — 17 February 2026 ⏭

PT-RAG: Structure-Fidelity Retrieval-Augmented Generation for Academic Papers

✇cs.AI, q-bio.NC updates on arXiv.org
By: Rui Yu · Tianyi Wang · Ruixia Liu · Yinglong Wang
17 February 2026 at 13:00
arXiv:2602.13647v1 Announce Type: cross Abstract: Retrieval-augmented generation (RAG) is increasingly applied to question-answering over long academic papers, where accurate evidence allocation under a fixed token budget is critical. Existing approaches typically flatten academic papers into unstructured chunks during preprocessing, which destroys the native hierarchical structure. This loss forces retrieval to operate in a disordered space, thereby producing fragmented contexts, misallocating tokens to non-evidential regions under finite token budgets, and increasing the reasoning burden for downstream language models. To address these issues, we propose PT-RAG, an RAG framework that treats the native hierarchical structure of academic papers as a low-entropy retrieval prior. PT-RAG first inherits the native hierarchy to construct a structure-fidelity PaperTree index, which prevents entropy increase at the source. It then designs a path-guided retrieval mechanism that aligns query semantics to relevant sections and selects high relevance root-to-leaf paths under a fixed token budget, yielding compact, coherent, and low-entropy retrieval contexts. In contrast to existing RAG approaches, PT-RAG avoids entropy increase caused by destructive preprocessing and provides a native low-entropy structural basis for subsequent retrieval. To assess this design, we introduce entropy-based structural diagnostics that quantify retrieval fragmentation and evidence allocation accuracy. On three academic question-answering benchmarks, PT-RAG achieves consistently lower section entropy and evidence alignment cross entropy than strong baselines, indicating reduced context fragmentation and more precise allocation to evidential regions. These structural advantages directly translate into higher answer quality.
There are no more articles
❌