cs.CL（2026-10-14）

📊 共 41 篇论文 | 🔗 7 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (26 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (13 🔗4) 支柱八：物理动画 (Physics-based Animation) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (26 篇)

#题目一句话要点标签🔗⭐
1 FACES: Fine-grained Gated Semantic Alignment for Multimodal Emotional Support Conversation 提出FACES框架以解决多模态情感支持对话中的语义反馈不足问题 multimodal chain-of-thought ✅
2 Are We Really Measuring Forgetting? Pitfalls in Evaluating Machine Unlearning for Multimodal LLMs 提出GEMMU以解决多模态机器遗忘评估问题 large language model multimodal
3 Recognizing Dialogue Ambiguity via Candidate Set Visual Grounding 提出候选集视觉定位以解决对话歧义问题 visual grounding
4 KoAGE: Do Large Language Models Know, Learn, and Coin Korean Neologisms? 提出KoAGE框架以评估语言模型对韩语新词的理解与生成能力 large language model
5 ChronoCulture: Evaluating Cultural Awareness Across Time in Large Language Models 提出ChronoCulture以评估大语言模型的文化意识随时间变化 large language model
6 OpenTWBench (Open Taiwan Benchmark): Taiwan-Domain Knowledge Evaluation for Large Language Models, and How the Measurement Apparatus Shapes the Score 提出OpenTWBench以评估大型语言模型的台湾领域知识 large language model
7 More Than Seeing: Vision as Scalable Anchors in Native Multimodal Pre-Training 提出视觉内容作为多模态学习的可扩展锚点 multimodal
8 AdaRSI: On-Demand Recursive Self-Improvement for Embodied Agents 提出AdaRSI框架以解决长时间任务中的自我改进问题 vision-language-action VLA
9 Cortex: Consolidating Context into Parametric Memory 提出Cortex以解决上下文整合与记忆保持问题 large language model multimodal
10 Frustratingly Easy Majority Voting Using Length Conditioning 提出长度条件多数投票方法以提升大语言模型推理准确性 large language model chain-of-thought ✅
11 SEA-HELM Audio: Evaluating Multilingual ASR-LLM Pipeline Performance Across SEA-HELM Subsets 提出SEA-HELM Audio评估框架以解决多语言ASR-LLM性能问题 large language model instruction following
12 HACHIMI-II: From Role-Play to Learner Behavior Modeling for Faithful Student Simulation 提出HACHIMI-II以解决学生模拟中的行为建模问题 large language model
13 Typed Decision Models on Tabular Data 提出类型决策模型以解决表格数据决策问题 large language model
14 PerSo: Investigating Facets of LLMs' Social Intelligence in Real-World Persuasion Scenarios 提出PerSo基准以评估LLMs在真实说服场景中的社交智能 large language model
15 CSWBench: Evaluating Context-Sensitive Code-Switching in Creative Writing 提出CSWBench以评估创作写作中的上下文敏感代码切换能力 large language model
16 Scaling Multilingual Safety and Reasoning Evaluation through Reasoning-Guided Translation 提出推理引导翻译框架以解决多语言安全性和推理评估问题 large language model
17 Which Knowledge Should LLMs Trust? Uncovering a Generalizable Steering Direction Across Knowledge Conflicts 提出一种可转移的知识选择方向以解决LLMs知识冲突问题 large language model
18 Interlock: From Local Correctness to Task Completion in Long-Horizon Agents 提出Interlock框架以解决长时间任务完成问题 large language model
19 SoK: Grading Open-Text Legal Answers, from Law School Exams to Trustworthy Rubric Judges 提出可信赖的评分标准以解决法律开放文本答案评估问题 large language model
20 MemDuet: Learning Dual-Process Memory for Long-Context Reasoning 提出MemDuet以解决长文档推理中的效率与性能问题 large language model
21 BANGLAPLAN: A Benchmark for Constraint-Based Planning and Decision-Making in Bangla 提出BANGLAPLAN以解决孟加拉语约束规划与决策评估问题 large language model ✅
22 Moral Combat: Exploiting Moral Foundations Theory to Jailbreak LLMs 提出多代理框架以利用道德基础理论破解大型语言模型 large language model
23 GraphRAG-LTM: A Layered Threat Modeling Framework for Attacks on Graph Retrieval-Augmented Generation 提出GraphRAG-LTM框架以解决图检索增强生成的安全性问题 large language model
24 Tail Competition Explains Scaling in Best-of-N Sampling for Verifiable Problems 提出尾部竞争机制以解释可验证问题中的Best-of-N采样扩展 large language model
25 LLM-jp-4 Technical Report 提出LLM-jp-4以推动日本主权大语言模型的发展 large language model
26 TopicMerge: A Divide and Conquer Approach to Topic Modeling 提出TopicMerge以解决主题建模中的效率与性能问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (13 篇)

#题目一句话要点标签🔗⭐
27 ReaDy: Recurrent States as Dynamic Context for Speculative Drafting in Hybrid Linear Attention Large Language Models 提出ReaDy以提升混合线性注意力模型的推理效率 linear attention large language model ✅
28 Rationale Distribution Distillation via Flow Matching for Distributional Continuous Latent Reasoning 提出基于流匹配的蒸馏方法以提升连续潜在推理能力 flow matching distillation chain-of-thought
29 PrismaCoder: Pushing the Boundaries of Open Multimodal Code Intelligence through Decomposed Rubric Synthesis 提出PrismaCoder以推动开放多模态代码智能的发展 reinforcement learning multimodal ✅
30 More Methods, Better Agents? A Reality Check on Self-Distillation in Agentic Reinforcement Learning 提出自蒸馏方法评估基准以解决强化学习训练不稳定问题 reinforcement learning distillation ✅
31 VASPO: Self-Distillation with Privileged Value Estimation 提出VASPO以提升自我蒸馏中的价值估计精度 PPO distillation privileged information ✅
32 Retrieve What Helps, Only When It Helps: Agentic Reinforcement Learning for Selective Retrieval-Augmented Machine Translation 提出选择性检索增强翻译以优化翻译质量 reinforcement learning large language model
33 Qwen-MT-Lumen at WMT2026: Instruction-Specialized Translation via Quality-Aware Training and On-Policy Distillation 提出Lumen以解决指令专用翻译的质量与遵循性问题 distillation instruction following
34 Beyond Corrected Actions: Training LLM Agents on Textual Critiques 提出CAFT框架以增强LLM代理的监督学习效果 reinforcement learning behavior cloning distillation
35 Manacá-1B-Instruct-2: Reproducible Safety Alignment for a Native Brazilian-Portuguese LLM. Negative DPO Results, a Controlled Composition Scan, and a Silent Failure Mode of the LLM Judge 提出Manacá-1B-Instruct-2以解决巴西葡萄牙语LLM安全对齐问题 DPO
36 Not All Criteria Matter: Importance-Aware Criteria Weighting for Rubric-Based LLM Evaluation 提出重要性感知标准加权方法以优化LLM评估 reinforcement learning large language model
37 Beyond State Capacity: Why Retrieval Breaks in Mamba 提出查询优先检索机制以解决Mamba模型的检索瓶颈问题 Mamba
38 PA1-OPD: One-Step Prefix-Augmented On-Policy Distillation for Language Model Pretraining 提出PA1-OPD以增强语言模型预训练中的知识蒸馏效果 distillation
39 ForkGRPO: Efficient Exploration and Fine-Grained Credit Assignment for Diffusion Language Models 提出ForkGRPO以解决扩散语言模型的探索与信用分配问题 reinforcement learning large language model

🔬 支柱八：物理动画 (Physics-based Animation) (2 篇)

#题目一句话要点标签🔗⭐
40 Taking the PULSE of Medical Reasoning: Step-Level Retrieval Timing for Knowledge-Grounded Inference 提出PUlSE以解决医学推理中的检索时机问题 PULSE
41 MakSafe: A Spatiotemporal Semantic Framework for Formal Specification and Compliance Checking of Road Vehicle Safety Rules 提出MakSafe框架以解决自动驾驶车辆安全规则验证问题 spatiotemporal

⬅️ 返回 cs.CL 首页 · 🏠 返回主页