2026-09论文中文摘要
GRRR：解码器LLM后训练中的重塑、旋转与路由几何
GRRR: The Geometry of Reshaping, Rotation, and Routing in Decoder LLM post-training
arXiv 2609.22146 · 2026-09-22弱连接，强信号：通过独立令牌采样实现扩散大语言模型的高效训练数据检测
Weak Ties, Strong Signals: Efficient Training Data Detection in Diffusion LLMs via Independent Token Sampling
arXiv 2609.22145 · 2026-09-22多语言安全信号是多层次的：过滤降低安全性的数据以构建更安全的LLM
Multilingual Safety Signals Are Multi-Layered: Filtering Safety-Degrading Data for Safer LLMs
arXiv 2609.22144 · 2026-09-22使用复合算子线性化LLM语义变换
Using Composition Operators to Linearize LLM Semantic Transformations
arXiv 2609.22143 · 2026-09-22学习何时优于启发式？Kubernetes调度器评分插件案例研究
When Does Learning Beat Heuristics? A Case Study in Kubernetes Scheduler Score Plugins
arXiv 2609.22142 · 2026-09-22注意力增强双分支ConvNeXt-BiLSTM网络用于受试者无关的脑电癫痫检测
Attention-Enhanced Dual-Branch ConvNeXt-BiLSTM Network for Subject-Independent EEG Seizure Detection
arXiv 2609.22141 · 2026-09-22ORCAS - 分布式声学传感的正交表示压缩
ORCAS - Orthogonal Representation for Compression of distributed Acoustic Sensing
arXiv 2609.22140 · 2026-09-22HFEMCNet：一种用于自动调制分类的紧凑型混合频率增强多通道网络
HFEMCNet: A Compact Hybrid Frequency Enriched Multi Channel Network for Automatic Modulation Classification
arXiv 2609.22139 · 2026-09-22真实性信号在语码混合下是否幸存？探测隐藏状态以检测印地英语码混合中的幻觉
Does the Truthfulness Signal Survive Code-Mixing? Probing Hidden States for Hallucination Detection in Hinglish
arXiv 2609.22138 · 2026-09-22流体天线信道重建与非负最小二乘检测器
Fluid Antenna Channel Reconstruction with Non-Negative Least Square Detector
arXiv 2609.22137 · 2026-09-22DiFA：用于词元级文本异常检测的双证据融合与聚合
DiFA: Dual Evidence Fusion and Aggregation for Token-Level Text Anomaly Detection
arXiv 2609.22136 · 2026-09-22读得最好并非操控最好：全模态大语言模型中的探测-操控层分离
Read-Best Is Not Steer-Best: A Probing--Steering Layer Dissociation in Omni-Modal Large Language Models
arXiv 2609.22135 · 2026-09-22用于频谱感知的低功耗超宽带接收机的实验评估
Experimental Evaluation of a Low-Power Ultra-Wideband Receiver for Spectrum Sensing
arXiv 2609.22134 · 2026-09-22LLM与人类标注的观测等价性
Observational Equivalence of LLM and Human Annotation
arXiv 2609.22133 · 2026-09-22WiNeRF：用于可操作无线信道建模的测量约束辐射场
WiNeRF: Measurement Constrained Radiance Fields for Actionable Wireless Channel Modeling
arXiv 2609.22132 · 2026-09-22面向大语言模型的关联感知结构化剪枝
Correlation-Aware Structured Pruning for Large Language Models
arXiv 2609.22131 · 2026-09-22基于层次贝叶斯优化的飞机多智能体系统之系统架构设计
Hierarchical Bayesian optimization of an aircraft-based multi-agent system-of-systems
arXiv 2609.22130 · 2026-09-22Helix-FNO：谱域算子学习与高保真机理模型耦合的快速代理仿真
Helix-FNO: Spectral-Domain Operator Learning Coupled with a High-Fidelity Mechanistic Model for Fast Surrogate Simulation
arXiv 2609.22129 · 2026-09-22ST-Topo GAN：一种匹配腕部运动复杂度的运动脑电到肌电解码模型
ST-Topo GAN: A Motor EEG-to-EMG Decoding Model Matched to Wrist Movement Complexity
arXiv 2609.22128 · 2026-09-22超越准确性与表面流畅性：面向法律条款生成的LLM风险敏感评估
Beyond Accuracy and Surface Fluency: Risk-Sensitive Evaluation of LLMs for Legal Clause Generation
arXiv 2609.22127 · 2026-09-22SolarFlowRefiner：面向地表太阳辐射降尺度的精化感知流匹配
SolarFlowRefiner: Refinement-Aware Flow Matching for Surface Solar Radiation Downscaling
arXiv 2609.22126 · 2026-09-22婆罗米系文字的类型驱动分词
Type-Driven Tokenization for Brahmic Scripts
arXiv 2609.22125 · 2026-09-22平衡RAG流水线中的推理与硬件约束：面向乌克兰多领域文档理解
Balancing Reasoning and Hardware Constraints in RAG Pipelines for Ukrainian Multi-Domain Document Understanding
arXiv 2609.22124 · 2026-09-22排名可移植性并不意味着可行性可移植性：联合硬件约束的目标特定评估
Rank Portability Does Not Imply Feasibility Portability: Target-Specific Evaluation of Joint Hardware Constraints
arXiv 2609.22122 · 2026-09-22基于深度表示学习的手机定位数据日常活动模式建模
Modelling daily activity patterns from mobile phone location data via deep representation learning
arXiv 2609.22121 · 2026-09-22成功留下绕行：为长时程智能体学习可执行的操作指南
Success Leaves Detours: Learning Executable Walkthroughs for Long-Horizon Agents
arXiv 2609.22120 · 2026-09-22评估感知随模型规模从格式转向上下文
Evaluation Awareness Shifts from Format to Context with Model Scale
arXiv 2609.22119 · 2026-09-22将商用射频感应发生器改装为计算机控制的真空与气体集成退火系统，用于活性金属晶粒生长
Retrofitting a commercial RF induction generator into a computer-controlled, vacuum and gas integrated annealing system for reactive-metal grain growth
arXiv 2609.22118 · 2026-09-22LE4Mob：面向人类移动建模的归纳式、距离感知与通用位置嵌入
LE4Mob: Towards Inductive, Distance-Aware and General-Purpose Location Embedding for Human Mobility Modelling
arXiv 2609.22117 · 2026-09-22基于LC谐振腔振铃衰减计数的对数ADC伽马/电子能谱仪
A Gamma/Electron Spectrometer with Logarithmic ADC based on LC Tank Ring-Down Oscillation Counting
arXiv 2609.22116 · 2026-09-22ZoAQ：通过查询重用耦合的自适应零阶查询
ZoAQ: Adaptive Zeroth-Order Querying via Query-Reuse Coupling
arXiv 2609.22115 · 2026-09-22多轮编码智能体中上下文压缩网关的经验成本归因
An Empirical Cost Attribution of Context-Compression Gateways in Multi-Turn Coding Agents
arXiv 2609.22114 · 2026-09-22面向阿片类药物使用障碍治疗保留与提前终止预测的机器学习模型公平性研究
Toward Fairness in Machine Learning Models for Predicting Treatment Retention and Premature Discontinuation in Medication for Opioid Use Disorder
arXiv 2609.22113 · 2026-09-22隐私个性化在LLMs中的权衡：风格测量信号减少对用户特定文本生成的影响
Privacy Personalization Trade offs in LLMs: The Impact of Stylometric Signal Reduction on User-Specific Text Generation
arXiv 2609.22112 · 2026-09-22超越文本：验证智能体撰写的论文是否得到其工件的支持
Beyond the Text: Verifying That Agent-Written Papers Are Backed by Their Artifacts
arXiv 2609.22111 · 2026-09-22评估面向非洲孕产妇和疫苗接种医疗场景的微调模型与基础语言模型
Evaluating Fine-Tuned and Base Language Models in Maternal and Vaccination Healthcare for African Settings
arXiv 2609.22110 · 2026-09-22共享学习率并非选择性在策略蒸馏中的中性控制
A Shared Learning Rate Is Not a Neutral Control in Selective On-Policy Distillation
arXiv 2609.22109 · 2026-09-22修正基于学习的感知以确保安全
Correcting Learning-based Perception for Safety
arXiv 2609.22108 · 2026-09-22通道缩减下最先进基础模型用于睡眠分析的比较研究
Comparative Analysis of State-of-the-Art Foundation Models for Sleep Analysis Under Channel Reduction
arXiv 2609.22105 · 2026-09-22DeepInstructor：一种用于经验驱动型想法评估的智能体式AI导师
DeepInstructor: An Agentic AI Instructor for Experience-Driven Idea Evaluation
arXiv 2609.22104 · 2026-09-22当你的身份能改变你获得的代码：LLM代码生成中人格诱导偏差的研究
When Who You Are Can Change the Code You Get: A Study of Persona-Induced Bias in LLM Code Generation
arXiv 2609.22102 · 2026-09-22上下文投毒作为长上下文语言模型中的极值注意力干扰
Context Poisoning as Extreme-Value Attention Interference in Long-Context Language Models
arXiv 2609.22101 · 2026-09-22AdaMem：检索增强生成中软压缩的自适应记忆令牌分配
AdaMem: Adaptive Memory Token Allocation for Soft Compression in Retrieval-Augmented Generation
arXiv 2609.22100 · 2026-09-22食谱数据结构框架及其在烹饪与营养洞察中的应用
A framework for recipe data structure with applications for culinary and nutritional insights
arXiv 2609.22099 · 2026-09-22TreeSpark：用于半自回归投机解码的校准、负载自适应草稿树
TreeSpark: Calibrated, Load-Adaptive Draft Trees for Semi-Autoregressive Speculative Decoding
arXiv 2609.22098 · 2026-09-22代码的Token签名：比较大型语言模型间的编码行为
Token Signatures of Code: Comparing Coding Behaviors Across Large Language Models
arXiv 2609.22097 · 2026-09-22AI推断的X平台上气候变化运动中的表达性幸福感与集体行动话语
AI-inferred expressed well-being and collective-action discourse in climate-change campaigns on X
arXiv 2609.22096 · 2026-09-22超越原始波形：融合EDA视觉表示用于压力检测
Beyond the Raw Waveform: Fusing Visual Representations of EDA for Stress Detection
arXiv 2609.22095 · 2026-09-22摘要、评判、优化：多模态内容审核的解耦内容理解与策略学习
Summarize, Judge, Refine: Decoupled Content Understanding and Policy Learning for Multimodal Content Moderation
arXiv 2609.22094 · 2026-09-22DC-CLM：面向AI数据中心动态的WECC复合负荷模型扩展
DC-CLM: Extending the WECC Composite Load Model for AI Data Center Dynamics
arXiv 2609.22093 · 2026-09-22
