cs.CL（2026-10-22）

📊 共 31 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (24 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 TRIAD: Benchmarking Omni-Modal Ambiguity in Multimodal Large Language Models 提出TRIAD基准以解决多模态大语言模型中的模糊性问题 large language model multimodal
2 Spectral Smoothing and Spiking Perturbations Reconfigure Capability Profiles and Selective Task Performance in Large Language Models 提出谱平滑与尖峰扰动以重构大语言模型能力特征 large language model chain-of-thought
3 The Geometry of Metaphor: Probing and Steering Metaphor Representations in Large Language Models 提出机制可解释性技术以控制大型语言模型中的隐喻表现 large language model
4 Meaning Contracts: Local Semantic Geometry in Large Language Models 提出局部语义几何以分析大语言模型中的意义契约 large language model
5 An Empirical Study of Recent Large Language Models for Knowledge Graph to Text Generation on Standard Benchmarks 基于大型语言模型的知识图谱到文本生成研究 large language model
6 Parameter-Efficient Adaptation of Large Language Models to a Low-Resource Agglutinative Language: A Corpus, Datasets, and a Reproducible Benchmark for the Bashkir Language 提出低资源语言Bashkir的高效适应方法以改善LLM表现 large language model
7 When LLMs Forget What They Believe: Characterizing State Integrity Collapse in Long-Horizon Dialogues 提出StateTrack-Bench以解决长对话中的状态完整性崩溃问题 large language model instruction following ✅
8 Synthetic-to-Real Code-Switched ASR: A Spontaneous Nepali-English Podcast Benchmark 提出合成到真实的代码切换ASR框架以解决尼泊尔语-英语识别问题 foundation model TAMP
9 When More Metadata Misleads: Probing Field-Level Grounding in LLM-Generated Museum Descriptions 提出元数据条件生成框架以解决博物馆描述生成问题 visual grounding
10 Prompting Strategy Effects Depend on the Judge: A Multilingual Study of LLM-as-a-Judge Robustness 提出多语言评估中的提示策略依赖于评判者的研究 chain-of-thought
11 Joint or Per-Option? Compact Encoders for Closed-Set Decisions from 17 Million to 1 Billion Parameters 提出紧凑编码器以优化闭集决策过程 large language model
12 ULME: Predictive Coupling for Mutually Reinforcing Embedding and Language Modeling 提出ULME框架以解决嵌入与语言建模的统一问题 large language model
13 Investigating Multi-Cultural Value Selection Biases in LLMs 研究多文化价值选择偏差以优化大型语言模型的文化理解 large language model
14 A Reality Check of Inference-Time Context Augmentation Methods 提出评估框架以检验推理时上下文增强方法的有效性 large language model
15 IQuest-Q1: Advancing Foundation Capabilities for Agentic CLI Systems 提出IQuest-Q1以提升智能CLI系统的基础能力 foundation model
16 Tracing Hallucinations in LLM Reasoning for UML Class Diagram Generation: An Empirical Taxonomy and Mitigation Study 提出一种分类法以解决LLM生成UML类图中的幻觉问题 large language model
17 Silicon Samples Match the Answers but Not the Respondents 提出响应者覆盖差距以评估合成样本的有效性 large language model
18 Compiling Autoregressive LLM Control into Policy-Exact Parallel Inference for Hierarchical RAG 提出层次化RAG的策略精确并行推理以优化自回归LLM控制 large language model
19 SAGE: Sense-Aligned Multimodal Grounding Evaluation with Gloss-Augmented Reasoning 提出SAGE以解决多模态基础评估问题 multimodal
20 AusClinBench: a deterministically scored benchmark of large language models against current Australian clinical policy 提出AusClinBench基准以评估澳大利亚临床政策下的语言模型 large language model
21 POS-Signature Demonstration Retrieval for In-Context Dependency Parsing 提出POS签名策略以优化上下文依赖解析示例检索 chain-of-thought
22 LazyFlash: Forward-Dense, Backward-Sparse Attention for Accelerated Long-Context Fine-Tuning 提出LazyFlash以加速长序列上下文微调 large language model
23 TaxoSAE: Rethinking Sparse Autoencoders through Taxonomic Logic 提出TaxoSAE以解决稀疏自编码器的知识组织问题 large language model
24 Omni-UBench: Benchmarking Omni-Modal LLM-based Guardrails Across Safety, Utility, and Reliability 提出Omni-UBench以评估多模态LLM安全性问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
25 Cola2: Co-Adaptive Hierarchical Language Modeling with Continuous Semantic Compression and Balanced Discrete Generation 提出Cola2框架以解决层次语言生成中的语义压缩与文本生成问题 flow matching representation learning classifier-free guidance
26 HERD: Harness-Evolving Recursive Distillation for Continual Learning 提出HERD框架以解决持续学习中的遗忘问题 distillation
27 Verbalizing Activation Differences with Natural Language Autoencoders 提出自然语言自编码器以有效表达激活差异 reinforcement learning large language model
28 ReflectPO: Reflection-Guided Policy Optimization for Long-Horizon LLM Agents 提出ReflectPO以解决长时间任务中的错误恢复问题 reinforcement learning large language model
29 Kolibri: A Sovereign European Model on the Pareto Frontier 提出Kolibri模型以支持德国及受监管领域的AI应用 reinforcement learning instruction following
30 Free Lunch for LLMs Confidence: The RL Critic Already Knows 提出利用现有RL评论员作为LLMs信心估计器以降低数据需求 reinforcement learning large language model

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
31 Scaling cross-situational word learning on children's naturalistic experiences 提出跨情境词汇学习模型以解决儿童词汇学习问题 egocentric

⬅️ 返回 cs.CL 首页 · 🏠 返回主页