cs.CL（2025-09-19）

📊 共 42 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (39 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (3)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (39 篇)

#题目一句话要点标签🔗⭐
1 Session-Level Spoken Language Assessment with a Multimodal Foundation Model via Multi-Target Learning 提出基于多模态基础模型和多目标学习的会话级口语评估方法 foundation model multimodal
2 RephQA: Evaluating Readability of Large Language Models in Public Health Question Answering RephQA：评估大型语言模型在公共健康问答中的可读性，并提出优化策略。 large language model chain-of-thought
3 Multi-Physics: A Comprehensive Benchmark for Multimodal LLMs Reasoning on Chinese Multi-Subject Physics Problems 提出Multi-Physics：一个用于评估多模态LLM在中文物理问题上推理能力的综合基准。 multimodal chain-of-thought ✅
4 Intrinsic Meets Extrinsic Fairness: Assessing the Downstream Impact of Bias Mitigation in Large Language Models 研究LLM内外部公平性：偏差缓解对下游任务的影响评估 large language model
5 Sparse-Autoencoder-Guided Internal Representation Unlearning for Large Language Models 提出基于稀疏自编码器的内部表征解学习方法，提升大语言模型的信息遗忘效果。 large language model
6 Concept Unlearning in Large Language Models via Self-Constructed Knowledge Triplets 提出基于自构建知识三元组的大语言模型概念遗忘方法 large language model
7 DivLogicEval: A Framework for Benchmarking Logical Reasoning Evaluation in Large Language Models DivLogicEval：用于评估大语言模型逻辑推理能力的新基准框架 large language model
8 CFDLLMBench: A Benchmark Suite for Evaluating Large Language Models in Computational Fluid Dynamics CFDLLMBench：用于评估大语言模型在计算流体动力学中应用能力的基准套件 large language model ✅
9 How to inject knowledge efficiently? Knowledge Infusion Scaling Law for Pre-training Large Language Models 提出知识注入缩放律，高效指导大语言模型预训练中的领域知识注入 large language model
10 Exploring Polyglot Harmony: On Multilingual Data Allocation for Large Language Models Pretraining 提出Climb框架，通过优化多语言数据分配提升大语言模型性能。 large language model
11 'Rich Dad, Poor Lad': How do Large Language Models Contextualize Socioeconomic Factors in College Admission ? 提出DPAF框架，揭示LLM在大学招生中对社会经济因素的偏见及推理机制 large language model
12 Longitudinal and Multimodal Recording System to Capture Real-World Patient-Clinician Conversations for AI and Encounter Research: Protocol 构建纵向多模态记录系统，捕捉真实医患对话，促进AI与诊疗研究 multimodal
13 The Role of High-Performance GPU Resources in Large Language Model Based Radiology Imaging Diagnosis 利用高性能GPU加速基于大语言模型的放射影像诊断 large language model
14 The Curious Case of Visual Grounding: Different Effects for Speech- and Text-based Language Encoders 研究视觉信息融入对语音和文本语言编码器内部表征的影响 visual grounding
15 Fine-Tuning Large Multimodal Models for Automatic Pronunciation Assessment 微调大型多模态模型用于自动发音评估，提升细粒度评估能力 multimodal
16 Can LLMs Judge Debates? Evaluating Non-Linear Reasoning via Argumentation Theory Semantics 评估LLM在辩论中的非线性推理能力：基于论证理论语义 large language model chain-of-thought
17 VOX-KRIKRI: Unifying Speech and Language through Continuous Fusion VOX-KRIKRI：通过连续融合统一语音和语言 large language model multimodal
18 BEFT: Bias-Efficient Fine-Tuning of Language Models BEFT：一种高效偏置项微调方法，提升语言模型在低数据场景下的性能。 large language model
19 Beyond the Score: Uncertainty-Calibrated LLMs for Automated Essay Assessment 提出基于不确定性校准的大语言模型，用于提升自动作文评分系统的可靠性。 large language model
20 Evaluating Behavioral Alignment in Conflict Dialogue: A Multi-Dimensional Comparison of LLM Agents and Humans 评估冲突对话中的行为对齐：LLM智能体与人类的多维度对比 large language model
21 RPG: A Repository Planning Graph for Unified and Scalable Codebase Generation 提出RPG，用于统一和可扩展的代码库生成 large language model
22 CultureScope: A Dimensional Lens for Probing Cultural Understanding in LLMs 提出 CultureScope，通过多维度文化知识分类评估LLM的文化理解能力。 large language model ✅
23 Think, Verbalize, then Speak: Bridging Complex Thoughts and Comprehensible Speech 提出Think-Verbalize-Speak框架，解耦推理与口语表达，提升口语对话系统性能。 large language model ✅
24 Re-FRAME the Meeting Summarization SCOPE: Fact-Based Summarization and Personalization via Questions 提出FRAME框架和SCOPE协议，解决会议摘要生成中幻觉、遗漏和个性化问题 large language model
25 Distribution-Aligned Decoding for Efficient LLM Task Adaptation 提出SVDecode，通过分布对齐解码高效适应LLM下游任务 large language model
26 Best-of-L: Cross-Lingual Reward Modeling for Mathematical Reasoning 提出Best-of-L跨语言奖励模型，提升多语言LLM在数学推理中的性能。 large language model
27 Once Upon a Time: Interactive Learning for Storytelling with Small Language Models 提出交互式学习框架，利用认知反馈提升小语言模型的故事生成能力。 large language model
28 Pipeline Parallelism is All You Need for Optimized Early-Exit Based Self-Speculative Decoding 提出流水线并行自推测解码PPSD，优化基于早退出的LLM推理。 large language model
29 LiteLong: Resource-Efficient Long-Context Data Synthesis for LLMs LiteLong：一种资源高效的长文本数据合成方法，用于训练大型语言模型 large language model
30 Computational Analysis of Conversation Dynamics through Participant Responsivity 通过参与者响应性分析对话动态，量化对话质量并区分不同对话。 large language model
31 Purely Semantic Indexing for LLM-based Generative Recommendation and Retrieval 提出纯语义索引，解决LLM生成式推荐与检索中的语义ID冲突问题 large language model
32 Evaluating the Effectiveness and Scalability of LLM-Based Data Augmentation for Retrieval 研究LLM数据增强在检索中的有效性和可扩展性，揭示最优增强策略。 large language model
33 Evaluating CxG Generalisation in LLMs via Construction-Based NLI Fine Tuning 提出ConTest-NLI基准，评估LLM在基于构式语法的NLI泛化能力 large language model
34 CodeRAG: Finding Relevant and Necessary Knowledge for Retrieval-Augmented Repository-Level Code Completion 提出CodeRAG以解决代码补全中的知识检索问题 large language model ✅
35 UniGist: Towards General and Hardware-aligned Sequence-level Long Context Compression UniGist：面向通用和硬件对齐的序列级长上下文压缩框架 large language model
36 REFER: Mitigating Bias in Opinion Summarisation via Frequency Framed Prompting 提出REFER框架，通过频率框架提示缓解意见摘要中的偏见 large language model
37 SciEvent: Benchmarking Multi-domain Scientific Event Extraction SciEvent：提出多领域科学事件抽取基准，促进结构化科学内容理解 large language model
38 DNA-DetectLLM: Unveiling AI-Generated Text via a DNA-Inspired Mutation-Repair Paradigm 提出DNA-DetectLLM，利用DNA修复机制实现零样本AI生成文本检测。 large language model ✅
39 A method for improving multilingual quality and diversity of instruction fine-tuning datasets 提出M-DaQ方法，提升多语言指令微调数据集的质量和多样性，增强LLM的多语言能力。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
40 Self-Rewarding Rubric-Based Reinforcement Learning for Open-Ended Reasoning 提出自奖励的基于规则的强化学习框架，提升开放式推理能力 reinforcement learning large language model
41 How do Language Models Generate Slang: A Systematic Comparison between Human and Machine-Generated Slang Usages 系统比较人类与LLM生成俚语用法，揭示LLM在俚语理解上的偏差 distillation large language model
42 It Depends: Resolving Referential Ambiguity in Minimal Contexts with Commonsense Knowledge 研究表明，大型语言模型在最小上下文中利用常识知识解决指代消解歧义方面存在困难。 direct preference optimization large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页