cs.CL（2026-03-10）

📊 共 19 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (18 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (18 篇)

#题目一句话要点标签🔗⭐
1 Do What I Say: A Spoken Prompt Dataset for Instruction-Following 提出DOWIS数据集，用于评估语音提示下语音大语言模型的性能。 large language model instruction following
2 Investigating Gender Stereotypes in Large Language Models via Social Determinants of Health 通过社会决定因素探究大型语言模型中的性别刻板印象 large language model
3 Model Merging in the Era of Large Language Models: Methods, Applications, and Future Directions 提出FUSE框架，综述大语言模型合并方法、应用与未来方向。 large language model
4 Benchmarking Political Persuasion Risks Across Frontier Large Language Models 评估前沿大语言模型在政治观点影响上的风险，发现模型间说服力存在显著差异。 large language model
5 Tracking Cancer Through Text: Longitudinal Extraction From Radiology Reports Using Open-Source Large Language Models 提出开放源代码模型以提取放射学报告中的肿瘤信息 large language model
6 CyberThreat-Eval: Can Large Language Models Automate Real-World Threat Research? 提出CyberThreat-Eval以解决现有CTI报告自动化不足问题 large language model ✅
7 DEO: Training-Free Direct Embedding Optimization for Negation-Aware Retrieval 提出DEO：一种免训练的直接嵌入优化方法，用于处理包含否定信息的检索任务。 large language model multimodal
8 You Didn't Have to Say It like That: Subliminal Learning from Faithful Paraphrases 通过语义不变的释义文本，语言模型可隐式学习教师模型的偏好。 chain-of-thought
9 RbtAct: Rebuttal as Supervision for Actionable Review Feedback Generation RbtAct：利用同行评审反驳意见，生成更具可操作性的评审反馈 large language model
10 ESAinsTOD: A Unified End-to-End Schema-Aware Instruction-Tuning Framework for Task-Oriented Dialog Modeling 提出ESAinsTOD，一个统一的、端到端的、模式感知的指令调优框架，用于任务型对话建模。 large language model
11 Common Sense vs. Morality: The Curious Case of Narrative Focus Bias in LLMs 揭示LLM叙事焦点偏差：道德推理优先于常识理解 large language model
12 One-Eval: An Agentic System for Automated and Traceable LLM Evaluation One-Eval：一个自动化、可溯源的 Agentic LLM 评估系统 large language model ✅
13 Beyond Fine-Tuning: Robust Food Entity Linking under Ontology Drift with FoodOntoRAG FoodOntoRAG：一种无需微调的、鲁棒的食品实体链接方法，可应对本体漂移。 large language model
14 Evaluation of LLMs in retrieving food and nutritional context for RAG systems 利用LLM驱动的RAG系统，高效检索食品营养数据，降低领域专家使用门槛。 large language model
15 ALARM: Audio-Language Alignment for Reasoning Models ALARM：通过音频-语言对齐增强推理模型的音频理解能力 chain-of-thought
16 Quantifying and extending the coverage of spatial categorization data sets 利用大型语言模型扩展空间范畴数据集，提升场景覆盖率 large language model
17 TA-Mem: Tool-Augmented Autonomous Memory Retrieval for LLM in Long-Term Conversational QA 提出TA-Mem框架，增强LLM在长期对话问答中自主记忆检索能力 large language model
18 Bioalignment: Measuring and Improving LLM Disposition Toward Biological Systems for AI Safety Bioalignment：通过微调提升LLM对生物系统的倾向性，增强AI安全性 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (1 篇)

#题目一句话要点标签🔗⭐
19 Reading, Not Thinking: Understanding and Bridging the Modality Gap When Text Becomes Pixels in Multimodal LLMs 针对多模态LLM中“文本像素化”的模态差异问题，提出自蒸馏方法显著提升视觉文本理解能力。 distillation large language model multimodal

⬅️ 返回 cs.CL 首页 · 🏠 返回主页