cs.CL（2026-02-09）

📊 共 26 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 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1)

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

#题目一句话要点标签🔗⭐
1 UReason: Benchmarking the Reasoning Paradox in Unified Multimodal Models 提出UReason基准测试，揭示统一多模态模型中推理对图像生成的影响悖论。 multimodal chain-of-thought
2 Pretraining with Token-Level Adaptive Latent Chain-of-Thought 提出Token级自适应潜在CoT预训练方法，提升语言模型效率与性能。 large language model chain-of-thought
3 Large Language Models for Geolocation Extraction in Humanitarian Crisis Response 提出基于大语言模型的地理定位提取框架，提升人道主义危机响应中文本地理信息提取的精度和公平性。 large language model
4 Large Language Models and Impossible Language Acquisition: "False Promise" or an Overturn of our Current Perspective towards AI 通过对比LLM与LSTM学习“不可能语言”的能力，反思AI发展范式 large language model
5 WorldTravel: A Realistic Multimodal Travel-Planning Benchmark with Tightly Coupled Constraints WorldTravel：一个具有紧耦合约束的真实多模态旅行规划基准 multimodal
6 When Does Context Help? Error Dynamics of Contextual Information in Large Language Models 提出Transformer大语言模型上下文信息影响的统一理论框架，分析误差动态。 large language model
7 Latent Reasoning with Supervised Thinking States 提出Thinking States，通过监督式思维状态实现高效的LLM潜在推理。 large language model chain-of-thought
8 Beyond Transcripts: A Renewed Perspective on Audio Chaptering 提出AudioSeg，一种基于音频表示的章节分割方法，显著优于文本方法。 multimodal instruction following
9 Affective Flow Language Model for Emotional Support Conversation 提出情感流语言模型AFlow，通过细粒度情感建模提升情感支持对话质量。 large language model ✅
10 Is Reasoning Capability Enough for Safety in Long-Context Language Models? 提出复合推理攻击模型以提升长文本语言模型的安全性 large language model
11 PERSPECTRA: A Scalable and Configurable Pluralist Benchmark of Perspectives from Arguments 提出PERSPECTRA以解决多元观点评估问题 large language model
12 Learning to Judge: LLMs Designing and Applying Evaluation Rubrics GER-Eval：探索LLM自主设计评估准则并应用于自然语言生成任务评估 large language model
13 Fundamental Reasoning Paradigms Induce Out-of-Domain Generalization in Language Models 通过诱导基本推理范式提升语言模型在域外任务的泛化能力 large language model
14 ValueFlow: Measuring the Propagation of Value Perturbations in Multi-Agent LLM Systems ValueFlow：用于评估多智能体LLM系统中价值扰动传播的框架 large language model
15 TEAM: Temporal-Spatial Consistency Guided Expert Activation for MoE Diffusion Language Model Acceleration TEAM：时序-空间一致性引导的专家激活加速MoE扩散语言模型 large language model ✅
16 Linguistics and Human Brain: A Perspective of Computational Neuroscience 计算神经科学视角下语言与人脑关系研究：利用深度学习弥合语言学与神经科学的鸿沟 large language model
17 LLMs and people both learn to form conventions -- just not with each other 研究表明，大型语言模型与人类均能形成对话惯例，但无法在人机交互中有效建立 multimodal

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
18 Improving Data and Reward Design for Scientific Reasoning in Large Language Models 提出Dr. SCI框架，提升大语言模型在开放式科学推理任务上的性能 reinforcement learning reward design large language model
19 VocalNet-MDM: Accelerating Streaming Speech LLM via Self-Distilled Masked Diffusion Modeling VocalNet-MDM：通过自蒸馏掩码扩散模型加速流式语音LLM distillation MDM large language model
20 Dynamic Long Context Reasoning over Compressed Memory via End-to-End Reinforcement Learning 提出基于强化学习的动态长文本推理框架，解决长文本处理中的效率和信息遗忘问题。 reinforcement learning large language model
21 Beyond Scalar Scores: Reinforcement Learning for Error-Aware Quality Estimation of Machine Translation 提出ALOPE-RL框架，利用强化学习和错误感知奖励提升机器翻译质量估计 reinforcement learning large language model
22 Document Reconstruction Unlocks Scalable Long-Context RLVR 提出基于文档重构的无监督RLVR方法，提升LLM长文本处理能力 reinforcement learning reward design large language model
23 WildReward: Learning Reward Models from In-the-Wild Human Interactions WildReward：从真实用户交互中学习奖励模型，提升LLM性能。 DPO large language model ✅
24 GISA: A Benchmark for General Information-Seeking Assistant GISA：通用信息搜索助手基准测试，解决现有基准测试不自然和数据污染问题。 imitation learning large language model
25 New Skills or Sharper Primitives? A Probabilistic Perspective on the Emergence of Reasoning in RLVR 提出概率框架以解释RLVR中推理能力的出现 reinforcement learning large language model

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
26 ViGoEmotions: A Benchmark Dataset For Fine-grained Emotion Detection on Vietnamese Texts 提出ViGoEmotions越南语细粒度情感检测数据集，并评估多种预训练模型。 motion prediction large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页