cs.CL（2025-09-20）

📊 共 21 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (17 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (17 篇)

#题目一句话要点标签🔗⭐
1 Decoding Uncertainty: The Impact of Decoding Strategies for Uncertainty Estimation in Large Language Models 对比搜索提升大语言模型不确定性估计的有效性 large language model
2 USB-Rec: An Effective Framework for Improving Conversational Recommendation Capability of Large Language Model 提出USB-Rec框架，提升大语言模型在对话推荐系统中的训练与推理能力 large language model
3 ConceptViz: A Visual Analytics Approach for Exploring Concepts in Large Language Models ConceptViz：一种用于探索大型语言模型概念的可视分析方法 large language model ✅
4 LLMsPark: A Benchmark for Evaluating Large Language Models in Strategic Gaming Contexts LLMsPark：提出基于博弈论的大语言模型战略能力评测基准 large language model ✅
5 The Sound of Syntax: Finetuning and Comprehensive Evaluation of Language Models for Speech Pathology 针对语音病理学，提出微调语言模型并进行全面评估，填补临床应用空白。 multimodal chain-of-thought
6 PruneCD: Contrasting Pruned Self Model to Improve Decoding Factuality 提出PruneCD，通过对比剪枝模型提升大型语言模型解码的事实性 large language model
7 Rethinking the Role of Text Complexity in Language Model Pretraining 研究文本复杂度对语言模型预训练的影响，揭示数据多样性与下游任务性能间的关系。 large language model
8 InteGround: On the Evaluation of Verification and Retrieval Planning in Integrative Grounding InteGround：提出综合性知识融合评估框架，用于评估LLM在复杂推理场景下的知识检索与验证能力。 large language model
9 AIPsychoBench: Understanding the Psychometric Differences between LLMs and Humans AIPsychoBench：构建LLM心理测量基准，揭示其与人类的差异及多语言影响 large language model
10 Assessing Classical Machine Learning and Transformer-based Approaches for Detecting AI-Generated Research Text 评估经典机器学习与Transformer模型在AI生成研究文本检测中的性能 large language model
11 Can an Individual Manipulate the Collective Decisions of Multi-Agents? M-Spoiler：利用单智能体知识攻击多智能体协同决策系统 large language model
12 The Oracle Has Spoken: A Multi-Aspect Evaluation of Dialogue in Pythia 通过多维度评估Pythia模型对话能力，揭示模型规模和微调的影响 large language model
13 Cognitive Linguistic Identity Fusion Score (CLIFS): A Scalable Cognition-Informed Approach to Quantifying Identity Fusion from Text 提出CLIFS，一种基于认知语言学和LLM的可扩展身份融合量化方法 large language model ✅
14 EG-MLA: Embedding-Gated Multi-head Latent Attention for Scalable and Efficient LLMs 提出EG-MLA，通过嵌入门控机制压缩KV缓存，提升LLM推理效率。 large language model
15 Robust Native Language Identification through Agentic Decomposition 提出基于Agent分解的NLI方法，提升模型在对抗性线索下的鲁棒性 large language model
16 Redefining Experts: Interpretable Decomposition of Language Models for Toxicity Mitigation 提出EigenShift方法，通过语言模型分解实现可解释的毒性内容抑制。 large language model
17 Challenging the Evaluator: LLM Sycophancy Under User Rebuttal 揭示LLM在用户反驳下的谄媚行为，警惕评估任务中的潜在风险 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🔗⭐
18 MCP: A Control-Theoretic Orchestration Framework for Synergistic Efficiency and Interpretability in Multimodal Large Language Models 提出基于模型-控制器-任务适配的MCP框架，提升多模态大模型的效率与可解释性。 reinforcement learning large language model multimodal
19 Reinforcement Learning Meets Large Language Models: A Survey of Advancements and Applications Across the LLM Lifecycle 综述：强化学习赋能大语言模型全生命周期，提升推理与对齐性能 reinforcement learning large language model
20 ChemOrch: Empowering LLMs with Chemical Intelligence via Synthetic Instructions ChemOrch：通过合成指令增强LLM的化学智能 distillation large language model

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
21 Semi-Supervised Synthetic Data Generation with Fine-Grained Relevance Control for Short Video Search Relevance Modeling 提出半监督合成数据生成方法，解决短视频搜索相关性建模中数据稀缺和细粒度相关性不足问题。 penetration

⬅️ 返回 cs.CL 首页 · 🏠 返回主页