cs.CL（2025-12-08）

📊 共 23 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) (18 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗1) 支柱一：机器人控制 (Robot Control) (1)

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

#题目一句话要点标签🔗⭐
1 DART: Leveraging Multi-Agent Disagreement for Tool Recruitment in Multimodal Reasoning DART：利用多智能体分歧，在多模态推理中进行工具选择 large language model multimodal
2 Bridging Code Graphs and Large Language Models for Better Code Understanding CGBridge：通过桥接代码图和大型语言模型提升代码理解能力 large language model instruction following
3 Complementary Learning Approach for Text Classification using Large Language Models 提出一种互补学习方法，利用大语言模型进行文本分类，兼顾成本效益与研究严谨性。 large language model chain-of-thought
4 Do Large Language Models Truly Understand Cross-cultural Differences? 提出SAGE基准，评估大语言模型在跨文化理解和推理方面的能力 large language model
5 When Large Language Models Do Not Work: Online Incivility Prediction through Graph Neural Networks 提出基于图神经网络的在线不文明行为预测方法，优于大型语言模型。 large language model
6 Investigating Training and Generalization in Faithful Self-Explanations of Large Language Models 通过持续学习提升大语言模型自解释的忠实性与泛化能力 large language model
7 NeSTR: A Neuro-Symbolic Abductive Framework for Temporal Reasoning in Large Language Models NeSTR：一种神经符号演绎框架，用于增强大语言模型的时间推理能力 large language model
8 HalluShift++: Bridging Language and Vision through Internal Representation Shifts for Hierarchical Hallucinations in MLLMs HalluShift++：通过内部表征偏移弥合语言与视觉，实现多模态大语言模型中的分层幻觉检测 large language model multimodal ✅
9 A Simple Method to Enhance Pre-trained Language Models with Speech Tokens for Classification 提出一种简单方法，利用语音token增强预训练语言模型，用于分类任务。 large language model multimodal ✅
10 Leveraging KV Similarity for Online Structured Pruning in LLMs 提出Token Filtering，利用KV相似性实现在线LLM结构化剪枝。 large language model
11 Segment, Embed, and Align: A Universal Recipe for Aligning Subtitles to Signing 提出SEA框架，用于通用且高效的手语视频字幕对齐 TAMP
12 Short-Context Dominance: How Much Local Context Natural Language Actually Needs? 研究表明，大型语言模型预测任务中，短语境通常已足够，并提出DaMCL指标检测长语境依赖，优化模型输出。 large language model
13 Do Generalisation Results Generalise? 研究表明大语言模型泛化能力评估结果在不同OOD数据集上不具备一致性。 large language model
14 Mary, the Cheeseburger-Eating Vegetarian: Do LLMs Recognize Incoherence in Narratives? 研究表明大型语言模型在识别叙事不连贯性方面存在局限性，尤其是在人物性格违背方面。 large language model
15 PCMind-2.1-Kaiyuan-2B Technical Report PCMind-2.1-Kaiyuan-2B：开源20亿参数模型，提升资源受限场景下的训练效率与效果。 large language model ✅
16 MoCoRP: Modeling Consistent Relations between Persona and Response for Persona-based Dialogue MoCoRP：提出建模Persona与Response一致性关系框架，提升Persona对话质量。 large language model ✅
17 SwissGov-RSD: A Human-annotated, Cross-lingual Benchmark for Token-level Recognition of Semantic Differences Between Related Documents 提出SwissGov-RSD跨语言基准数据集，用于识别相关文档间语义差异。 large language model
18 Beyond Real: Imaginary Extension of Rotary Position Embeddings for Long-Context LLMs 扩展旋转位置编码RoPE的虚部，提升长文本LLM的建模能力 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
19 Enhancing Agentic RL with Progressive Reward Shaping and Value-based Sampling Policy Optimization 提出PRS和VSPO，提升Agentic RL在工具集成推理任务中的性能与泛化性 reinforcement learning PPO reward design
20 Adaptation of Embedding Models to Financial Filings via LLM Distillation 提出一种基于LLM蒸馏的金融文档嵌入模型自适应方法，提升金融领域信息检索性能。 distillation large language model
21 Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning 提出原生并行推理器NPR，通过自蒸馏强化学习赋能LLM并行推理能力 reinforcement learning large language model
22 Persian-Phi: Efficient Cross-Lingual Adaptation of Compact LLMs via Curriculum Learning 提出Persian-Phi，通过课程学习高效地将小型LLM跨语言适配到波斯语 curriculum learning large language model ✅

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
23 On the Interplay of Pre-Training, Mid-Training, and RL on Reasoning Language Models 提出可控实验框架，解析预训练、中期训练和强化学习对推理语言模型的影响 manipulation reinforcement learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页