语言大模型 / LLM - arXivDaily 专题

2509.24637 2025-10-14 cs.SE 67%

Bridging Developer Instructions and Code Completion Through Instruction-Aware Fill-in-the-Middle Paradigm

Zhensu Sun, Chengran Yang, Chao Peng, Pengfei Gao, Xiaoning Du, Li Li, David Lo

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2510.09781 2025-10-14 cs.LG cs.AI cs.CL 67%

Building a Foundational Guardrail for General Agentic Systems via Synthetic Data

Yue Huang, Hang Hua, Yujun Zhou, Pengcheng Jing, Manish Nagireddy, Inkit Padhi, Greta Dolcetti, Zhangchen Xu, Subhajit Chaudhury, Ambrish Rawat, Liubov Nedoshivina, Pin-Yu Chen, Prasanna Sattigeri, Xiangliang Zhang

机构 * University of Notre Dame（诺特大学）； MIT-IBM Watson AI Lab（MIT-IBM Watson AI实验室）； University of Washington（华盛顿大学）； Ca’ Foscari University of Venice（威尼斯卡福尔学院）； IBM Research（IBM研究院）

专题命中预训练与数据：LLM(abstract)；分类 cs.CL、cs.AI、cs.LG

2501.15857 2025-10-14 cs.AI cs.CL cs.LG 67%

Are Transformers Able to Reason by Connecting Separated Knowledge in Training Data?

Yutong Yin, Zhaoran Wang

专题命中预训练与数据：prompting(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted by ICLR 2025

2510.09031 2025-10-13 cs.SI 67%

Web Crawler Restrictions, AI Training Datasets \& Political Biases

Paul Bouchaud, Pedro Ramaciotti

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2510.08485 2025-10-10 cs.CV 67%

InstructX: Towards Unified Visual Editing with MLLM Guidance

Chong Mou, Qichao Sun, Yanze Wu, Pengze Zhang, Xinghui Li, Fulong Ye, Songtao Zhao, Qian He

机构 * Intelligent Creation Team, ByteDance（字节跳动智能创作团队）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2510.08404 2025-10-10 cs.CL cs.AI cs.LG 67%

Single layer tiny Co$^4$ outpaces GPT-2 and GPT-BERT

Noor Ul Zain, Mohsin Raza, Ahsan Adeel

机构 * CMI-Lab University of Stirling（CMI实验室斯特林大学）

专题命中预训练与数据：pretraining(abstract)；分类 cs.CL、cs.AI、cs.LG

2510.06590 2025-10-09 cs.CV 67%

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer

Ziyuan Huang, DanDan Zheng, Cheng Zou, Rui Liu, Xiaolong Wang, Kaixiang Ji, Weilong Chai, Jianxin Sun, Libin Wang, Yongjie Lv, Taozhi Huang, Jiajia Liu, Qingpei Guo, Ming Yang, Jingdong Chen, Jun Zhou

机构 * Inclusion AI Ant Group（Inclusion AI Ant集团）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

Comments Code released at https://github.com/inclusionAI/Ming-UniVision

2510.00606 2025-10-09 cs.DC 67%

ElasWave: An Elastic-Native System for Scalable Hybrid-Parallel Training

Xueze Kang, Guangyu Xiang, Yuxin Wang, Hao Zhang, Yuchu Fang, Yuhang Zhou, Zhenheng Tang, Youhui Lv, Eliran Maman, Mark Wasserman, Alon Zameret, Zhipeng Bian, Shushu Chen, Zhiyou Yu, Jin Wang, Xiaoyu Wu, Yang Zheng, Chen Tian, Xiaowen Chu

专题命中预训练与数据：LLM(abstract)；pretraining(abstract)

2510.01860 2025-10-03 eess.AS cs.SD 67%

SLAP: Learning Speaker and Health-Related Representations from Natural Language Supervision

Angelika Ando, Auguste Crabeil, Adrien Lesage, Rachid Riad

专题命中预训练与数据：foundation model(abstract)；pretraining(abstract)

2407.20177 2025-10-03 cs.LG cs.AI cs.CL stat.ML 67%

AutoScale: Scale-Aware Data Mixing for Pre-Training LLMs

Feiyang Kang, Yifan Sun, Bingbing Wen, Si Chen, Dawn Song, Rafid Mahmood, Ruoxi Jia

机构 * Virginia Tech（弗吉尼亚理工大学）； UIUC（伊利诺伊大学香槟分校）； University of Washington（华盛顿大学）； UC Berkeley（加州大学伯克利分校）； University of Ottawa（渥太华大学）； NVIDIA（英伟达）

专题命中预训练与数据：LLM(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Published as a conference paper at COLM 2025

2510.00993 2025-10-02 cs.CV 67%

Visual Self-Refinement for Autoregressive Models

Jiamian Wang, Ziqi Zhou, Chaithanya Kumar Mummadi, Sohail Dianat, Majid Rabbani, Raghuveer Rao, Chen Qiu, Zhiqiang Tao

机构 * Rochester Institute of Technology（罗切斯特理工大学）； Bosch Research（博世研究）； DEVCOM Army Research Laboratory（美国陆军研究实验室）

专题命中预训练与数据：language model(abstract)；pretraining(abstract)

Comments Accepted by EMNLP2025

2509.25144 2025-09-30 cs.CL cs.AI cs.LG 67%

Paired by the Teacher: Turning Unpaired Data into High-Fidelity Pairs for Low-Resource Text Generation

Yen-Ju Lu, Thomas Thebaud, Laureano Moro-Velazquez, Najim Dehak, Jesus Villalba

机构 * Center for Language and Speech Processing, Johns Hopkins University（语言与语音处理中心，约翰霍普金斯大学）

专题命中预训练与数据：LLM(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted at EMNLP 2025 (Main Conference)

2509.24791 2025-09-30 cs.CV 67%

Vision Function Layer in Multimodal LLMs

Cheng Shi, Yizhou Yu, Sibei Yang

机构 * Sun Yat-sen University（中山大学）； School of Computing and Data Science（计算与数据科学学院）； The University of Hong Kong（香港大学）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

Comments Accepted at NeurIPS 2025 (preview; camera-ready in preparation)

2509.19677 2025-09-25 cs.CR 67%

Unmasking Fake Careers: Detecting Machine-Generated Career Trajectories via Multi-layer Heterogeneous Graphs

Michiharu Yamashita, Thanh Tran, Delvin Ce Zhang, Dongwon Lee

专题命中预训练与数据：large language model(abstract)；language model(abstract)

Comments Accepted at EMNLP 2025 Main

2411.11098 2025-09-22 cs.CV 67%

MolParser: End-to-end Visual Recognition of Molecule Structures in the Wild

Xi Fang, Jiankun Wang, Xiaochen Cai, Shangqian Chen, Shuwen Yang, Haoyi Tao, Nan Wang, Lin Yao, Linfeng Zhang, Guolin Ke

机构 * DP Technology（DP技术）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2509.14008 2025-09-18 cs.CL cs.AI cs.LG 67%

Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale

Hasan Abed Al Kader Hammoud, Mohammad Zbeeb, Bernard Ghanem

专题命中预训练与数据：language model(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Technical Report

2509.12893 2025-09-17 cs.CV 67%

MEJO: MLLM-Engaged Surgical Triplet Recognition via Inter- and Intra-Task Joint Optimization

Yiyi Zhang, Yuchen Yuan, Ying Zheng, Jialun Pei, Jinpeng Li, Zheng Li, Pheng-Ann Heng

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2509.03336 2025-09-04 q-bio.MN cs.IR q-bio.QM 67%

AI-Driven Drug Repurposing through miRNA-mRNA Relation

Sharanya Manoharan, Balu Bhasuran, Oviya Ramalakshmi Iyyappan, Mohamed Saleem Abdul Shukkoor, Malathi Sellapan, Kalpana Raja

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2509.01658 2025-09-03 cs.RO 67%

MoTo: A Zero-shot Plug-in Interaction-aware Navigation for General Mobile Manipulation

Zhenyu Wu, Angyuan Ma, Xiuwei Xu, Hang Yin, Yinan Liang, Ziwei Wang, Jiwen Lu, Haibin Yan

专题命中预训练与数据：language model(abstract)；foundation model(abstract)

Comments Accepted to CoRL 2025. Project Page: https://gary3410.github.io/MoTo/

2509.01644 2025-09-03 cs.CV 67%

OpenVision 2: A Family of Generative Pretrained Visual Encoders for Multimodal Learning

Yanqing Liu, Xianhang Li, Letian Zhang, Zirui Wang, Zeyu Zheng, Yuyin Zhou, Cihang Xie

机构 * University of California Santa Cruz（加州大学圣克鲁兹分校）； Apple（苹果公司）； University of California Berkeley（加州大学伯克利分校）

专题命中预训练与数据：foundation model(abstract)；pretraining(abstract)

2509.00613 2025-09-03 eess.IV cs.CV 67%

Promptable Longitudinal Lesion Segmentation in Whole-Body CT

Yannick Kirchhoff, Maximilian Rokuss, Fabian Isensee, Klaus H. Maier-Hein

机构 * German Cancer Research Center (DKFZ) Heidelberg（德国癌症研究中心（DKFZ）海德堡）； Division of Medical Image Computing（医学影像计算 division）； HIDSS4Health - Helmholtz Information and Data Science School for Health（HIDSS4Health - 赫尔姆霍茨信息与数据科学健康学校）； Faculty of Mathematics and Computer Science（数学与计算机科学学院）； Heidelberg University（海德堡大学）； Helmholtz Imaging（赫尔姆霍茨成像）； Pattern Analysis and Learning Group（模式分析与学习小组）； Department of Radiation Oncology（放射肿瘤科）

专题命中预训练与数据：pretraining(abstract)；prompting(abstract)

2509.00083 2025-09-03 cs.LG cs.AI cs.CL 67%

Data Cartography for Detecting Memorization Hotspots and Guiding Data Interventions in Generative Models

Laksh Patel, Neel Shanbhag

机构 * Illinois Mathematics and Science Academy（伊利诺伊数学与科学学院）

专题命中预训练与数据：pretraining(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments 6 pages, 2 figures, 1 table; Presented at the 42nd International Conference on Machine Learning (ICML), winning the "Best Poster" award at ICML's workshop for data in generative models (DIG-BUGS)

2508.04141 2025-08-29 eess.AS cs.SD 67%

Parallel GPT: Harmonizing the Independence and Interdependence of Acoustic and Semantic Information for Zero-Shot Text-to-Speech

Jingyuan Xing, Zhipeng Li, Jialong Mai, Xiaofen Xing, Xiangmin Xu

机构 * School of Future Technology, South China University of Technology（未来技术学院，华南理工大学）； School of Electronic and Information, South China University of Technology（电子与信息学院，华南理工大学）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

Comments Submitted to IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP)

2504.13203 2025-08-26 cs.CR cs.AI cs.CL cs.LG cs.MA 67%

X-Teaming: Multi-Turn Jailbreaks and Defenses with Adaptive Multi-Agents

Salman Rahman, Liwei Jiang, James Shiffer, Genglin Liu, Sheriff Issaka, Md Rizwan Parvez, Hamid Palangi, Kai-Wei Chang, Yejin Choi, Saadia Gabriel

专题命中预训练与数据：language model(abstract)；分类 cs.CL、cs.AI、cs.LG

2508.15404 2025-08-25 cs.CV 67%

From Linearity to Non-Linearity: How Masked Autoencoders Capture Spatial Correlations

Anthony Bisulco, Rahul Ramesh, Randall Balestriero, Pratik Chaudhari

机构 * University of Pennsylvania（宾夕法尼亚大学）； Brown University（布朗大学）

专题命中预训练与数据：foundation model(abstract)；pretraining(abstract)

2507.06484 2025-08-21 cs.GR cs.CV 67%

3D-Generalist: Self-Improving Vision-Language-Action Models for Crafting 3D Worlds

Fan-Yun Sun, Shengguang Wu, Christian Jacobsen, Thomas Yim, Haoming Zou, Alex Zook, Shangru Li, Yu-Hsin Chou, Ethem Can, Xunlei Wu, Clemens Eppner, Valts Blukis, Jonathan Tremblay, Jiajun Wu, Stan Birchfield, Nick Haber

机构 * Stanford University（斯坦福大学）； Nvidia Research（英伟达研究）

专题命中预训练与数据：foundation model(abstract)；pretraining(abstract)

Comments project website: https://ai.stanford.edu/~sunfanyun/3d-generalist/

2508.05580 2025-08-08 cs.CV 67%

Follow-Your-Instruction: A Comprehensive MLLM Agent for World Data Synthesis

Kunyu Feng, Yue Ma, Xinhua Zhang, Boshi Liu, Yikuang Yuluo, Yinhan Zhang, Runtao Liu, Hongyu Liu, Zhiyuan Qin, Shanhui Mo, Qifeng Chen, Zeyu Wang

机构 * HKUST(GZ)（香港科技大学（广州））； HKUST（香港科技大学）； Tsinghua Univerisity（清华大学）； Peking University（北京大学）； Chongqing University（重庆大学）； Beijing Innovation Center of Humanoid Robotics（北京人形机器人创新中心）

专题命中预训练与数据：large language model(abstract)；language model(abstract)

2412.13395 2025-08-05 cs.CL cs.AI cs.HC cs.LG 67%

Enhancing Talk Moves Analysis in Mathematics Tutoring through Classroom Teaching Discourse

Jie Cao, Abhijit Suresh, Jennifer Jacobs, Charis Clevenger, Amanda Howard, Chelsea Brown, Brent Milne, Tom Fischaber, Tamara Sumner, James H. Martin

机构 * School of Computer Science, University of Oklahoma（俄克拉荷马大学计算机科学学院）； Institute of Cognitive Science, University of Colorado Boulder（科罗拉多大学博尔德分校认知科学研究所）

专题命中预训练与数据：pretraining(abstract)；分类 cs.CL、cs.AI、cs.LG

Comments Accepted to COLING'2025

2508.01153 2025-08-05 cs.CV 67%

TEACH: Text Encoding as Curriculum Hints for Scene Text Recognition

Xiahan Yang, Hui Zheng

专题命中预训练与数据：language model(abstract)；pretraining(abstract)

Comments 9 pages (w/o ref), 5 figures, 7 tables

2507.15542 2025-08-05 cs.CV 67%

HOLa: Zero-Shot HOI Detection with Low-Rank Decomposed VLM Feature Adaptation

Qinqian Lei, Bo Wang, Robby T. Tan

机构 * National University of Singapore（国立新加坡大学）； University of Mississippi（密苏里大学）； ASUS Intelligent Cloud Services (AICS)（ASUS智能云服务（AICS））

专题命中预训练与数据：LLM(abstract)；language model(abstract)

Comments Accepted by ICCV 2025

1. 预训练与数据 12486 篇

Bridging Developer Instructions and Code Completion Through Instruction-Aware Fill-in-the-Middle Paradigm

Building a Foundational Guardrail for General Agentic Systems via Synthetic Data

Are Transformers Able to Reason by Connecting Separated Knowledge in Training Data?

Web Crawler Restrictions, AI Training Datasets \&amp; Political Biases

InstructX: Towards Unified Visual Editing with MLLM Guidance

Single layer tiny Co$^4$ outpaces GPT-2 and GPT-BERT

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer

ElasWave: An Elastic-Native System for Scalable Hybrid-Parallel Training

SLAP: Learning Speaker and Health-Related Representations from Natural Language Supervision

AutoScale: Scale-Aware Data Mixing for Pre-Training LLMs

Visual Self-Refinement for Autoregressive Models

Paired by the Teacher: Turning Unpaired Data into High-Fidelity Pairs for Low-Resource Text Generation

Vision Function Layer in Multimodal LLMs

Unmasking Fake Careers: Detecting Machine-Generated Career Trajectories via Multi-layer Heterogeneous Graphs

MolParser: End-to-end Visual Recognition of Molecule Structures in the Wild

Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale

MEJO: MLLM-Engaged Surgical Triplet Recognition via Inter- and Intra-Task Joint Optimization

AI-Driven Drug Repurposing through miRNA-mRNA Relation

MoTo: A Zero-shot Plug-in Interaction-aware Navigation for General Mobile Manipulation

OpenVision 2: A Family of Generative Pretrained Visual Encoders for Multimodal Learning

Promptable Longitudinal Lesion Segmentation in Whole-Body CT

Data Cartography for Detecting Memorization Hotspots and Guiding Data Interventions in Generative Models

Parallel GPT: Harmonizing the Independence and Interdependence of Acoustic and Semantic Information for Zero-Shot Text-to-Speech

X-Teaming: Multi-Turn Jailbreaks and Defenses with Adaptive Multi-Agents

From Linearity to Non-Linearity: How Masked Autoencoders Capture Spatial Correlations

3D-Generalist: Self-Improving Vision-Language-Action Models for Crafting 3D Worlds

Follow-Your-Instruction: A Comprehensive MLLM Agent for World Data Synthesis

Enhancing Talk Moves Analysis in Mathematics Tutoring through Classroom Teaching Discourse

TEACH: Text Encoding as Curriculum Hints for Scene Text Recognition

HOLa: Zero-Shot HOI Detection with Low-Rank Decomposed VLM Feature Adaptation

Web Crawler Restrictions, AI Training Datasets \& Political Biases