cs.CL（2025-12-24）

📊 共 20 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) (13) 支柱二：RL算法与架构 (RL & Architecture) (5 🔗1) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱四：生成式动作 (Generative Motion) (1)

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

#	题目	一句话要点	标签	🔗	⭐
1	Morality is Contextual: Learning Interpretable Moral Contexts from Human Data with Probabilistic Clustering and Large Language Models	COMETH框架通过概率聚类和LLM学习可解释的道德上下文，提升道德判断准确性。	large language model
2	Semi-Supervised Learning for Large Language Models Safety and Content Moderation	提出半监督学习方法，提升大语言模型安全与内容审核能力	large language model
3	Neural Probe-Based Hallucination Detection for Large Language Models	提出基于神经探针的大语言模型幻觉检测框架，提升低误报下的检测精度。	large language model
4	ClarifyMT-Bench: Benchmarking and Improving Multi-Turn Clarification for Conversational Large Language Models	提出ClarifyMT-Bench，用于评测和提升会话大语言模型的多轮澄清能力。	large language model
5	Foundation Model-based Evaluation of Neuropsychiatric Disorders: A Lifespan-Inclusive, Multi-Modal, and Multi-Lingual Study	提出基于大模型的神经精神疾病评估框架FEND，实现多模态、多语言和全生命周期的诊断。	foundation model
6	Evaluating Novelty in AI-Generated Research Plans Using Multi-Workflow LLM Pipelines	利用多工作流LLM评估AI生成研究计划的新颖性	large language model multimodal
7	Rethinking Supervised Fine-Tuning: Emphasizing Key Answer Tokens for Improved LLM Accuracy	SFTKey：通过强化关键答案token优化LLM监督微调的准确性	large language model chain-of-thought
8	Reflection Pretraining Enables Token-Level Self-Correction in Biological Sequence Models	提出反射预训练，使生物序列模型具备token级自纠错能力	large language model chain-of-thought
9	ReaSeq: Unleashing World Knowledge via Reasoning for Sequential Modeling	ReaSeq：通过推理释放世界知识，用于序列建模，提升推荐系统性能。	large language model chain-of-thought
10	Teaching People LLM's Errors and Getting it Right	研究LLM错误模式教学，提升用户识别LLM失效场景的能力	large language model
11	C2LLM Technical Report: A New Frontier in Code Retrieval via Adaptive Cross-Attention Pooling	C2LLM：通过自适应跨注意力池化实现代码检索的新突破	large language model
12	LLM_annotate: A Python package for annotating and analyzing fiction characters	LLM_annotate：用于小说人物分析的Python工具包，提升标注和分析效率。	large language model
13	Architectural Trade-offs in Small Language Models Under Compute Constraints	研究计算约束下小型语言模型架构权衡，揭示不同架构和训练预算对性能的影响	large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#	题目	一句话要点	标签	🔗	⭐
14	Distilling the Essence: Efficient Reasoning Distillation via Sequence Truncation	提出基于序列截断的推理蒸馏方法，提升小模型推理效率	distillation chain-of-thought	✅
15	MultiMind at SemEval-2025 Task 7: Crosslingual Fact-Checked Claim Retrieval via Multi-Source Alignment	TriAligner：通过多源对齐实现跨语言的事实验证声明检索	representation learning contrastive learning large language model
16	Where Did This Sentence Come From? Tracing Provenance in LLM Reasoning Distillation	提出推理蒸馏溯源框架，分析学生模型能力来源并指导数据选择。	teacher-student distillation
17	Semantic Refinement with LLMs for Graph Representations	提出DAS框架，利用LLM进行图表示语义增强，解决图结构异构性问题	representation learning large language model
18	NVIDIA Nemotron 3: Efficient and Open Intelligence	NVIDIA Nemotron 3：高效开放的智能模型家族，支持百万token上下文	reinforcement learning Mamba

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

#	题目	一句话要点	标签	🔗	⭐
19	Decoding Predictive Inference in Visual Language Processing via Spatiotemporal Neural Coherence	提出基于时空神经相干性的视觉语言处理预测推理解码框架	optical flow spatiotemporal multimodal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#	题目	一句话要点	标签	🔗	⭐
20	Optimizing Decoding Paths in Masked Diffusion Models by Quantifying Uncertainty	通过量化不确定性优化掩蔽扩散模型的解码路径	MDM

⬅️ 返回 cs.CL 首页 · 🏠 返回主页