cs.CL（2025-02-03）

📊 共 20 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) (13 🔗2) 支柱一：机器人控制 (Robot Control) (3) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (13 篇)

#题目一句话要点标签🔗⭐
1 Evaluation of Large Language Models via Coupled Token Generation 提出基于耦合Token生成的大语言模型评估方法，减少评估样本量并发现传统评估的潜在偏差。 large language model
2 Large Language Models Are Human-Like Internally 大型语言模型内部机制更贴近人类认知过程，优于以往认知建模研究结论 large language model
3 What is a Number, That a Large Language Model May Know It? 揭示大语言模型中数字表示的字符串与数值双重性及影响 large language model
4 AlignVLM: Bridging Vision and Language Latent Spaces for Multimodal Document Understanding AlignVLM：通过对齐视觉和语言隐空间，提升多模态文档理解能力。 multimodal
5 OphthBench: A Comprehensive Benchmark for Evaluating Large Language Models in Chinese Ophthalmology OphthBench：构建中文眼科领域LLM综合评测基准，助力临床应用 large language model
6 On Bob Dylan: A Computational Perspective 利用计算方法分析鲍勃·迪伦歌词，揭示其音乐风格的演变与创新 large language model
7 Lifelong Knowledge Editing requires Better Regularization 提出MPES与范数约束正则化方法，解决终身知识编辑中的模型退化问题 large language model
8 LLM-TA: An LLM-Enhanced Thematic Analysis Pipeline for Transcripts from Parents of Children with Congenital Heart Disease 提出LLM-TA：一种LLM增强的主题分析流程，用于分析先天性心脏病患儿父母的访谈记录。 large language model ✅
9 FutureVision: A methodology for the investigation of future cognition 提出FutureVision方法，结合多模态语义分析与眼动追踪，研究未来认知。 multimodal
10 Massive Values in Self-Attention Modules are the Key to Contextual Knowledge Understanding 发现自注意力模块中显著值是上下文知识理解的关键 large language model ✅
11 Plan-Then-Execute: An Empirical Study of User Trust and Team Performance When Using LLM Agents As A Daily Assistant 研究LLM智能体作为日常助手时，用户信任与团队表现的影响，采用Plan-Then-Execute模式。 large language model
12 COVE: COntext and VEracity prediction for out-of-context images COVE：通过上下文预测和真实性验证解决脱离语境的图像误导问题 multimodal
13 Jailbreaking with Universal Multi-Prompts 提出JUMP：一种利用通用多提示词破解大型语言模型的方法 large language model

🔬 支柱一：机器人控制 (Robot Control) (3 篇)

#题目一句话要点标签🔗⭐
14 FALCON: Fine-grained Activation Manipulation by Contrastive Orthogonal Unalignment for Large Language Model FALCON：基于对比正交解耦的大语言模型细粒度激活操控 manipulation large language model
15 Topic-FlipRAG: Topic-Orientated Adversarial Opinion Manipulation Attacks to Retrieval-Augmented Generation Models 提出Topic-FlipRAG，针对RAG模型进行主题导向的对抗性观点操纵攻击 manipulation large language model
16 Bias Beware: The Impact of Cognitive Biases on LLM-Driven Product Recommendations 利用认知偏差对抗LLM产品推荐：揭示模型脆弱性与操纵风险 manipulation 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🔗⭐
17 SelfCheck-Eval: A Multi-Module Framework for Zero-Resource Hallucination Detection in Large Language Models 提出SelfCheck-Eval框架，用于零资源检测大语言模型中的幻觉问题，并构建数学推理幻觉数据集AIME。 preference learning large language model
18 ReGLA: Refining Gated Linear Attention ReGLA：通过优化门控线性注意力机制提升大语言模型性能 linear attention large language model
19 Adaptive Distraction: Probing LLM Contextual Robustness with Automated Tree Search 提出基于树搜索的自适应干扰生成框架，提升LLM上下文鲁棒性压力测试效率。 DPO large language model ✅

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
20 Latent Lexical Projection in Large Language Models: A Novel Approach to Implicit Representation Refinement 提出潜在词汇投影（LLP）方法，提升大型语言模型词汇表征和文本生成质量。 implicit representation large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页