语言大模型 / LLM - arXivDaily 专题

2602.00088 2026-02-03 cs.LG cs.AI 85%

From Numbers to Prompts: A Cognitive Symbolic Transition Mechanism for Lightweight Time-Series Forecasting

从数字到提示：一种轻量时间序列预测的认知符号转换机制

Namkyung Yoon, Hwangnam Kim

机构 * School of Electrical Engineering, Korea University（韩国大学电气工程学院）

专题命中效率与部署：large language model(abstract)；language model(abstract)；foundation model(abstract)；small language model(abstract)

AI总结本文提出STM机制，通过符号抽象和提示工程实现高效时间序列预测，显著提升模型效率并降低资源消耗。

Comments 16 pages, 5 figures. Submitted to ACM Transactions on Intelligent Systems and Technology

2601.21611 2026-01-30 cs.IR cs.AI cs.CL 85%

Thinking Broad, Acting Fast: Latent Reasoning Distillation from Multi-Perspective Chain-of-Thought for E-Commerce Relevance

深度思考，快速行动：多视角链式推理的潜在推理蒸馏用于电子商务相关性

Baopu Qiu, Hao Chen, Yuanrong Wu, Changtong Zan, Chao Wei, Weiru Zhang, Xiaoyi Zeng

机构 * Alibaba International Digital Commerce Group（阿里巴巴国际数字商业集团）； Zhejiang University（浙江大学）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；SFT(abstract)

AI总结本文提出多视角链式推理蒸馏方法，通过改进的教师模型和轻量级学生模型提升电子商务搜索相关性建模的准确性和效率。

Comments 12 pages, 6 figures, Accepted by WWW2026 industry track

2508.14904 2026-01-21 cs.CL cs.AI 85%

Efficient Switchable Safety Control in LLMs via Magic-Token-Guided Co-Training

通过魔法令牌引导的协同训练实现LLM中的高效可切换安全控制

Jianfeng Si, Lin Sun, Zhewen Tan, Xiangzheng Zhang

机构 * Qiyuan Tech, Beijing, China（北京奇元科技）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；SFT(abstract)

AI总结本文提出通过魔法令牌引导的协同训练框架，实现LLM内容安全的高效可切换控制，提升安全性能并降低训练与部署成本。

Comments 15 pages,3 figures,5 tables

2601.09694 2026-01-15 cs.CL cs.AI cs.CV 85%

LLMs can Compress LLMs: Adaptive Pruning by Agents

LLMs可以压缩LLMs：通过代理的自适应剪枝

Sai Varun Kodathala, Rakesh Vunnam

机构 * Research and Development, Sports Vision, Inc.（研发部，Sports Vision公司）； Research and Development, Vizworld Inc.（研发部，Vizworld公司）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；foundation model(abstract)

AI总结本文提出通过代理引导的自适应剪枝方法，有效压缩LLMs，提升事实知识保留和模型性能。

Comments 17 Pages

2505.15683 2026-01-05 cs.CL cs.AI cs.DC 85%

FedSEA-LLaMA: A Secure, Efficient and Adaptive Federated Splitting Framework for Large Language Models

FedSEA-LLaMA: 一种安全、高效且适应性的联邦分割框架用于大语言模型

Zishuai Zhang, Hainan zhang, Weihua Li, Qinnan zhang, jin Dong, Yongxin Tong, Zhiming Zheng

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.AI

AI总结 FedSEA-LLaMA通过注入噪声、压缩注意力掩码和动态调整分割点，实现安全高效的联邦分割学习，提升大语言模型的训练和推断速度。

2512.22309 2025-12-30 cs.LG cs.AI 85%

LLMBoost: Make Large Language Models Stronger with Boosting

LLMBoost: 通过提升使大型语言模型更强

Zehao Chen, Tianxiang Ai, Yifei Li, Gongxun Li, Yuyang Wei, Wang Zhou, Guanghui Li, Bin Yu, Zhijun Chen, Hailong Sun, Fuzhen Zhuang, Jianxin Li, Deqing Wang, Yikun Ban

机构 * Beihang University（北航大学）； China Telecom eSurfing Cloud（中国电信云）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

AI总结 LLMBoost通过引入跨模型注意力机制、链式训练范式和近似并行推理范式，提升大型语言模型的性能和效率。

2510.23649 2025-12-24 cs.LG cs.AI 85%

Efficient Low Rank Attention for Long-Context Inference in Large Language Models

高效低秩注意力用于大语言模型长上下文推断

Tenghui Li, Guoxu Zhou, Xuyang Zhao, Yuning Qiu, Qibin Zhao

机构 * Guangdong University of Technology（广东工业大学）； RIKEN AIP（理化学研究所AIP）； Key Laboratory of Intelligent Detection and the Internet of Things in Manufacturing, Ministry of Education, Guangzhou, CHINA（教育部智能制造智能化检测与物联网重点实验室）； RIKEN iTHEMS（理化学研究所iTHEMS）； RIKEN IMS（理化学研究所IMS）； Chiba University（千叶大学）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

AI总结 LRQK通过低秩分解和混合缓存机制，实现大语言模型长上下文推断中的高效低秩注意力，减少内存使用并保持高精度。

Comments https://neurips.cc/virtual/2025/loc/san-diego/poster/118451

Journal ref 39th Conference on Neural Information Processing Systems (NeurIPS 2025)

2512.07454 2025-12-09 cs.CL cs.AI 85%

Persian-Phi: Efficient Cross-Lingual Adaptation of Compact LLMs via Curriculum Learning

Persian-Phi: 通过课程学习高效适应紧凑大语言模型

Amir Mohammad Akhlaghi, Amirhossein Shabani, Mostafa Abdolmaleki, Saeed Reza Kheradpisheh

机构 * Shahid Beheshti University（沙希德·贝赫谢普大学）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；instruction tuning(abstract)

AI总结 Persian-Phi通过课程学习高效适应紧凑大语言模型，实现波斯语多语言能力的扩展。

2512.05073 2025-12-05 cs.LG cs.AI cs.AR cs.SE 85%

David vs. Goliath: Can Small Models Win Big with Agentic AI in Hardware Design?

大卫与歌利亚：小型模型能否通过代理AI在硬件设计中胜出？

Shashwat Shankar, Subhranshu Pandey, Innocent Dengkhw Mochahari, Bhabesh Mali, Animesh Basak Chowdhury, Sukanta Bhattacharjee, Chandan Karfa

机构 * Indian Institute of Technology, Guwahati, India（印度理工学院冈瓦提分校）； NXP USA, Inc.（NXP美国公司）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；foundation model(abstract)

AI总结本研究探讨小型语言模型结合代理AI在硬件设计中的应用，证明其在成本效率和性能上的优势。

2510.09942 2025-10-14 cs.LG cs.AI cs.IT math.IT 85%

Conformal Sparsification for Bandwidth-Efficient Edge-Cloud Speculative Decoding

Payel Bhattacharjee, Fengwei Tian, Meiyu Zhong, Guangyi Zhang, Osvaldo Simeone, Ravi Tandon

机构 * Department of Electrical and Computer Engineering（电气与计算机工程系；亚利桑那大学；美国亚利桑那州图森）； University of Arizona（信息科学与电子工程学院；浙江大学；中国杭州310027）； Tucson, AZ, USA（工程系；伦敦国王学院；英国伦敦）； College of Information Science and Electronic Engineering, Zhejiang University, Hangzhou 310027, China ； Department of Engineering ； King’s College London ； London, UK

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；small language model(abstract)

Comments 39th Conference on Neural Information Processing Systems (NeurIPS 2025) Workshop: AI and ML for Next-Generation Wireless Communications and Networking (AI4NextG)

2510.09722 2025-10-14 cs.CL cs.AI cs.CV 85%

Layout-Aware Parsing Meets Efficient LLMs: A Unified, Scalable Framework for Resume Information Extraction and Evaluation

Fanwei Zhu, Jinke Yu, Zulong Chen, Ying Zhou, Junhao Ji, Zhibo Yang, Yuxue Zhang, Haoyuan Hu, Zhenghao Liu

机构 * Hangzhou City University（杭州市大学）； Alibaba Group（阿里巴巴集团）； Zhejiang Lab（浙江实验室）； Alibaba Cloud（阿里云）； Northeastern University（东北大学）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；instruction tuning(abstract)

2510.05197 2025-10-08 cs.AI cs.LG stat.AP stat.ML 85%

Efficient Prediction of Pass@k Scaling in Large Language Models

Joshua Kazdan, Rylan Schaeffer, Youssef Allouah, Colin Sullivan, Kyssen Yu, Noam Levi, Sanmi Koyejo

机构 * Stanford University（斯坦福大学）； University of Toronto（多伦多大学）； EPFL（苏黎世联邦理工学院）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

2509.25803 2025-10-01 cs.IR cs.AI cs.CE cs.LG 85%

Better with Less: Small Proprietary Models Surpass Large Language Models in Financial Transaction Understanding

Wanying Ding, Savinay Narendra, Xiran Shi, Adwait Ratnaparkhi, Chengrui Yang, Nikoo Sabzevar, Ziyan Yin

机构 * JPMorgan Chase & Co.（摩根大通公司）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

Comments 9 pages, 5 figures

2509.05263 2025-09-09 cs.AI cs.CV cs.LG 85%

LatticeWorld: A Multimodal Large Language Model-Empowered Framework for Interactive Complex World Generation

Yinglin Duan, Zhengxia Zou, Tongwei Gu, Wei Jia, Zhan Zhao, Luyi Xu, Xinzhu Liu, Yenan Lin, Hao Jiang, Kang Chen, Shuang Qiu

机构 * NetEase, Inc.（网易公司）； Beihang University（北京航空航天大学）； Tsinghua University（清华大学）； City University of Hong Kong（香港城市大学）； Independent Researcher & Technical Artists（独立研究者及技术艺术家）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

2501.16658 2025-08-11 cs.CL cs.AI 85%

Contextual Reinforcement in Multimodal Token Compression for Large Language Models

Naderdel Piero, Zacharias Cromwell, Nathaniel Wainwright, Matthias Nethercott

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.AI

Comments arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship

2508.01625 2025-08-05 cs.LG cs.AI 85%

EAC-MoE: Expert-Selection Aware Compressor for Mixture-of-Experts Large Language Models

Yuanteng Chen, Yuantian Shao, Peisong Wang, Jian Cheng

机构 * Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）； Nanjing University of Science and Technology（南京理工大学）； AIRIA

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

Comments 22 pages, 13 figures. ACL 2025

2506.13817 2025-06-18 q-bio.GN cs.AI cs.LG cs.SE q-bio.QM 85%

DeepSeq: High-Throughput Single-Cell RNA Sequencing Data Labeling via Web Search-Augmented Agentic Generative AI Foundation Models

Saleem A. Al Dajani, Abel Sanchez, John R. Williams

机构 * Massachusetts Institute of Technology（麻省理工学院）

专题命中效率与部署：foundation model(title,abstract)；large language model(comments,journal_ref)；language model(comments,journal_ref)；分类 cs.AI、cs.LG

Comments 4 pages, 5 figures, Accepted by ICML 2025 FM4LS https://openreview.net/forum?id=zNjXOZxEYB . Workshop on Multi-modal Foundation Models and Large Language Models for Life Sciences (FM4LS)}, July 2025

Journal ref International Conference on Machine Learning (ICML). Workshop on Multi-modal Foundation Models and Large Language Models for Life Sciences (FM4LS), July 2025

2503.02969 2025-06-17 cs.CL cs.AI 85%

InfiniSST: Simultaneous Translation of Unbounded Speech with Large Language Model

Siqi Ouyang, Xi Xu, Lei Li

机构 * Carnegie Mellon University（卡内基梅隆大学）； Language Technologies Institute（语言技术研究所）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.AI

Comments ACL 2025 Findings

2408.08696 2025-05-27 cs.CL cs.LG 85%

Turning Trash into Treasure: Accelerating Inference of Large Language Models with Token Recycling

Xianzhen Luo, Yixuan Wang, Qingfu Zhu, Zhiming Zhang, Xuanyu Zhang, Qing Yang, Dongliang Xu

机构 * Harbin Institute of Technology（哈尔滨工业大学）； Du Xiaoman (Beijing) Science Technology Co., Ltd.（杜小满（北京）科技有限公司）

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.LG

Comments Accepted by ACL2025. Code is [here](https://github.com/Luowaterbi/TokenRecycling). Token Recycling has already merged into [SpecBench](https://github.com/hemingkx/Spec-Bench)

2505.15774 2025-05-22 cs.CL cs.LG 85%

Beyond Hard and Soft: Hybrid Context Compression for Balancing Local and Global Information Retention

Huanxuan Liao, Wen Hu, Yao Xu, Shizhu He, Jun Zhao, Kang Liu

机构 * Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； University of Chinese Academy of Sciences（中国科学院大学）； Ant Group（蚂蚁集团）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；instruction tuning(abstract)

2505.07345 2025-05-13 cs.CL cs.AI cs.IR 85%

QUPID: Quantified Understanding for Enhanced Performance, Insights, and Decisions in Korean Search Engines

Ohjoon Kwon, Changsu Lee, Jihye Back, Lim Sun Suk, Inho Kang, Donghyeon Jeon

机构 * Naver Corporation（纳维尔公司）

专题命中效率与部署：LLM(abstract)；large language model(abstract)；language model(abstract)；small language model(abstract)

Journal ref ACL 2025 Industry Track

2505.02362 2025-05-06 cs.CR cs.AI cs.LG 85%

Advancing Email Spam Detection: Leveraging Zero-Shot Learning and Large Language Models

Ghazaleh SHirvani, Saeid Ghasemshirazi

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

2405.13828 2025-04-21 cs.CL cs.AI 85%

Babysit A Language Model From Scratch: Interactive Language Learning by Trials and Demonstrations

Ziqiao Ma, Zekun Wang, Joyce Chai

专题命中效率与部署：language model(title,abstract)；large language model(abstract,comments)；分类 cs.CL、cs.AI

Comments NAACL 2025 (Main) & Workshop on Large Language Models and Cognition @ ICML 2024 (Oral)

2504.06446 2025-04-10 cs.LG cs.AI 85%

Can you Finetune your Binoculars? Embedding Text Watermarks into the Weights of Large Language Models

Fay Elhassan, Niccolò Ajroldi, Antonio Orvieto, Jonas Geiping

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

2412.13335 2025-04-08 cs.CL cs.AI 85%

Training Dynamics of a 1.7B LLaMa Model: A Data-Efficient Approach

Miles Q. Li, Benjamin C. M. Fung, Shih-Chia Huang

专题命中效率与部署：large language model(abstract)；language model(abstract)；instruction tuning(abstract)；pretraining(abstract)

2501.13999 2025-03-26 cs.CL cs.AI 85%

Framework for Progressive Knowledge Fusion in Large Language Models Through Structured Conceptual Redundancy Analysis

Joseph Sakau, Evander Kozlowski, Roderick Thistledown, Basil Steinberger

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.AI

Comments arXiv admin note: This paper has been withdrawn by arXiv due to disputed and unverifiable authorship

2502.00046 2025-02-04 cs.LG cs.CL 85%

Optimization Strategies for Enhancing Resource Efficiency in Transformers & Large Language Models

Tom Wallace, Naser Ezzati-Jivan, Beatrice Ombuki-Berman

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.LG

Comments Accepted for ACM's ICPE 2025 in Short Paper format

2402.05894 2024-10-28 cs.AI cs.LG 85%

Large Language Model Meets Graph Neural Network in Knowledge Distillation

Shengxiang Hu, Guobing Zou, Song Yang, Yanglan Gan, Bofeng Zhang, Yixin Chen

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

Comments This work has been submitted to the IEEE for possible publication

2410.14766 2024-10-22 cs.SE cs.AI cs.ET cs.LG cs.PL 85%

Evaluating Quantized Large Language Models for Code Generation on Low-Resource Language Benchmarks

Enkhbold Nyamsuren

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.AI、cs.LG

2410.00558 2024-10-02 cs.CL cs.AI cs.SE 85%

AMR-Evol: Adaptive Modular Response Evolution Elicits Better Knowledge Distillation for Large Language Models in Code Generation

Ziyang Luo, Xin Li, Hongzhan Lin, Jing Ma, Lidong Bing

专题命中效率与部署：large language model(title)；language model(title)；分类 cs.CL、cs.AI

Comments EMNLP 2024

1. 效率与部署 22368 篇

From Numbers to Prompts: A Cognitive Symbolic Transition Mechanism for Lightweight Time-Series Forecasting

Thinking Broad, Acting Fast: Latent Reasoning Distillation from Multi-Perspective Chain-of-Thought for E-Commerce Relevance

Efficient Switchable Safety Control in LLMs via Magic-Token-Guided Co-Training

LLMs can Compress LLMs: Adaptive Pruning by Agents

FedSEA-LLaMA: A Secure, Efficient and Adaptive Federated Splitting Framework for Large Language Models

LLMBoost: Make Large Language Models Stronger with Boosting

Efficient Low Rank Attention for Long-Context Inference in Large Language Models

Persian-Phi: Efficient Cross-Lingual Adaptation of Compact LLMs via Curriculum Learning

David vs. Goliath: Can Small Models Win Big with Agentic AI in Hardware Design?

Conformal Sparsification for Bandwidth-Efficient Edge-Cloud Speculative Decoding

Layout-Aware Parsing Meets Efficient LLMs: A Unified, Scalable Framework for Resume Information Extraction and Evaluation

Efficient Prediction of Pass@k Scaling in Large Language Models

Better with Less: Small Proprietary Models Surpass Large Language Models in Financial Transaction Understanding

LatticeWorld: A Multimodal Large Language Model-Empowered Framework for Interactive Complex World Generation

Contextual Reinforcement in Multimodal Token Compression for Large Language Models

EAC-MoE: Expert-Selection Aware Compressor for Mixture-of-Experts Large Language Models

DeepSeq: High-Throughput Single-Cell RNA Sequencing Data Labeling via Web Search-Augmented Agentic Generative AI Foundation Models

InfiniSST: Simultaneous Translation of Unbounded Speech with Large Language Model

Turning Trash into Treasure: Accelerating Inference of Large Language Models with Token Recycling

Beyond Hard and Soft: Hybrid Context Compression for Balancing Local and Global Information Retention

QUPID: Quantified Understanding for Enhanced Performance, Insights, and Decisions in Korean Search Engines

Advancing Email Spam Detection: Leveraging Zero-Shot Learning and Large Language Models

Babysit A Language Model From Scratch: Interactive Language Learning by Trials and Demonstrations

Can you Finetune your Binoculars? Embedding Text Watermarks into the Weights of Large Language Models

Training Dynamics of a 1.7B LLaMa Model: A Data-Efficient Approach

Framework for Progressive Knowledge Fusion in Large Language Models Through Structured Conceptual Redundancy Analysis

Optimization Strategies for Enhancing Resource Efficiency in Transformers & Large Language Models

Large Language Model Meets Graph Neural Network in Knowledge Distillation

Evaluating Quantized Large Language Models for Code Generation on Low-Resource Language Benchmarks

AMR-Evol: Adaptive Modular Response Evolution Elicits Better Knowledge Distillation for Large Language Models in Code Generation