cs.CL（2025-08-21）

📊 共 36 篇论文 | 🔗 11 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (27 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗4) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (27 篇)

#题目一句话要点标签🔗⭐
1 Unveiling Trust in Multimodal Large Language Models: Evaluation, Analysis, and Mitigation 提出MultiTrust-X基准，用于评估、分析和缓解多模态大语言模型中的信任问题。 large language model multimodal chain-of-thought
2 EcomMMMU: Strategic Utilization of Visuals for Robust Multimodal E-commerce Models EcomMMMU：面向电商多模态模型的视觉信息策略性利用 large language model multimodal ✅
3 DeepMEL: A Multi-Agent Collaboration Framework for Multimodal Entity Linking DeepMEL：提出一种多智能体协作框架，用于解决多模态实体链接任务。 large language model multimodal
4 WangchanThaiInstruct: An instruction-following Dataset for Culture-Aware, Multitask, and Multi-domain Evaluation in Thai WangchanThaiInstruct：一个用于评估泰语文化感知、多任务和多领域的大型语言模型数据集。 large language model instruction following
5 ContextualLVLM-Agent: A Holistic Framework for Multi-Turn Visually-Grounded Dialogue and Complex Instruction Following 提出CoLVLM-Agent框架，解决多轮视觉对话和复杂指令跟随中的上下文理解难题。 large language model instruction following
6 Confidence-Modulated Speculative Decoding for Large Language Models 提出基于置信度调制的推测解码方法，加速大语言模型的自回归推理。 large language model
7 EMNLP: Educator-role Moral and Normative Large Language Models Profiling EMNLP：构建教育者角色道德规范大语言模型评估框架，评估伦理风险 large language model ✅
8 Dream 7B: Diffusion Large Language Models Dream 7B：提出一种基于扩散的更强大的开放域大语言模型 large language model
9 Attribution, Citation, and Quotation: A Survey of Evidence-based Text Generation with Large Language Models 综述论文：对大型语言模型中基于证据的文本生成方法进行归因、引用和引用的研究 large language model
10 A Survey on Large Language Model Benchmarks 系统性综述大语言模型评测基准，分析现有问题并提出未来设计方向。 large language model
11 Self-Guided Function Calling in Large Language Models via Stepwise Experience Recall 提出SEER，通过逐步经验回忆实现大语言模型中的自引导函数调用 large language model
12 Dissecting Tool-Integrated Reasoning: An Empirical Study and Analysis 提出ReasonZoo基准，剖析工具集成推理对大语言模型推理能力的提升 large language model chain-of-thought
13 Retrieval-Augmented Review Generation for Poisoning Recommender Systems 提出RAGAN框架，通过检索增强生成高质量评论，提升推荐系统数据投毒攻击的有效性和隐蔽性。 foundation model multimodal
14 Annif at the GermEval-2025 LLMs4Subjects Task: Traditional XMTC Augmented by Efficient LLMs Annif系统在GermEval-2025 LLMs4Subjects任务中，通过高效LLM增强传统XMTC方法，获得第一名。 large language model
15 Exploiting Vocabulary Frequency Imbalance in Language Model Pre-training 研究词汇频率不平衡对语言模型预训练的影响，揭示扩大词表主要降低高频词的不确定性。 large language model
16 SparK: Query-Aware Unstructured Sparsity with Recoverable KV Cache Channel Pruning SparK：查询感知的非结构化稀疏化KV缓存通道剪枝，提升长文本LLM推理效率。 large language model ✅
17 SurGE: A Benchmark and Evaluation Framework for Scientific Survey Generation SurGE：用于科学综述生成的基准测试与评估框架 large language model ✅
18 Trained Miniatures: Low cost, High Efficacy SLMs for Sales & Marketing 提出“训练微型模型”，以低成本高效地进行销售和营销领域的文本生成。 large language model
19 Subjective Behaviors and Preferences in LLM: Language of Browsing 提出HeTLM以更好地捕捉用户的主观行为与偏好 large language model
20 When Audio and Text Disagree: Revealing Text Bias in Large Audio-Language Models 提出MCR-BENCH基准，揭示大型音频语言模型在不一致多模态输入中存在的文本偏见问题。 multimodal ✅
21 Conflict-Aware Soft Prompting for Retrieval-Augmented Generation 提出CARE模型，通过冲突感知软提示缓解RAG中的上下文-记忆冲突问题 large language model
22 SemToken: Semantic-Aware Tokenization for Efficient Long-Context Language Modeling SemToken：面向长文本高效建模的语义感知分词方法 large language model
23 Dancing with Deer: A Constructional Perspective on MWEs in the Era of LLMs 利用构式语法视角，研究大型语言模型时代下多词表达的理解与泛化。 large language model
24 SafetyFlow: An Agent-Flow System for Automated LLM Safety Benchmarking 提出SafetyFlow，一个全自动Agent-Flow系统，用于大规模语言模型（LLM）安全基准测试。 large language model
25 AmbiSQL: Interactive Ambiguity Detection and Resolution for Text-to-SQL AmbiSQL：交互式歧义检测与消解，提升Text-to-SQL准确率 large language model ✅
26 Are Checklists Really Useful for Automatic Evaluation of Generative Tasks? 研究检查列表在生成任务自动评估中的有效性，并提出选择性使用策略。 large language model ✅
27 Identifying and Answering Questions with False Assumptions: An Interpretable Approach 提出一种可解释的方法，用于识别并回答带有错误假设的问题。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
28 CARFT: Boosting LLM Reasoning via Contrastive Learning with Annotated Chain-of-Thought-based Reinforced Fine-Tuning CARFT：通过对比学习与基于CoT的强化微调提升LLM推理能力 reinforcement learning contrastive learning large language model ✅
29 Fin-PRM: A Domain-Specialized Process Reward Model for Financial Reasoning in Large Language Models 提出Fin-PRM：一种领域专用过程奖励模型，用于提升大语言模型在金融推理中的能力。 reinforcement learning distillation large language model ✅
30 SDGO: Self-Discrimination-Guided Optimization for Consistent Safety in Large Language Models 提出SDGO，利用自判别引导优化提升大语言模型安全性 reinforcement learning large language model ✅
31 Do Cognitively Interpretable Reasoning Traces Improve LLM Performance? 研究表明，认知可解释的推理轨迹不一定能提升LLM在开放域问答任务中的性能。 distillation large language model chain-of-thought
32 End-to-End Agentic RAG System Training for Traceable Diagnostic Reasoning 提出Deep-DxSearch，一种端到端Agentic RAG系统，用于可追溯的诊断推理。 reinforcement learning reward design large language model ✅
33 Influence-driven Curriculum Learning for Pre-training on Limited Data 提出基于训练数据影响力的课程学习方法，提升有限数据下预训练语言模型性能。 curriculum learning
34 Position Bias Mitigates Position Bias:Mitigate Position Bias Through Inter-Position Knowledge Distillation 提出Pos2Distill框架，通过位置间知识蒸馏缓解长文本理解中的位置偏差问题 distillation
35 SLM4Offer: Personalized Marketing Offer Generation Using Contrastive Learning Based Fine-Tuning SLM4Offer：基于对比学习微调的个性化营销方案生成模型 contrastive learning

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
36 Adversarial Attacks against Neural Ranking Models via In-Context Learning 提出Few-Shot Adversarial Prompting (FSAP)，利用上下文学习攻击神经排序模型。 manipulation large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页