cs.CL（2025-12-04）

📊 共 22 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) (18 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗1)

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

#题目一句话要点标签🔗⭐
1 Enhancing Clinical Note Generation with ICD-10, Clinical Ontology Knowledge Graphs, and Chain-of-Thought Prompting Using GPT-4 利用ICD-10、临床知识图谱和CoT提示，增强GPT-4的临床笔记生成能力 large language model chain-of-thought
2 Challenging the Abilities of Large Language Models in Italian: a Community Initiative CALAMITA：意大利语大型语言模型能力评测的社区驱动基准 large language model
3 Model Whisper: Steering Vectors Unlock Large Language Models' Potential in Test-time 提出测试时引导向量，无需微调即可提升大语言模型在特定任务上的性能。 large language model
4 LexGenius: An Expert-Level Benchmark for Large Language Models in Legal General Intelligence 提出LexGenius：一个专家级中文法律通用智能大语言模型评测基准。 large language model ✅
5 RapidUn: Influence-Driven Parameter Reweighting for Efficient Large Language Model Unlearning RapidUn：基于影响力的参数重加权高效实现大语言模型遗忘 large language model
6 To Think or Not to Think: The Hidden Cost of Meta-Training with Excessive CoT Examples CoT-Recipe：通过调节CoT样本比例提升元训练中LLM的推理能力 large language model chain-of-thought
7 Arbitrage: Efficient Reasoning via Advantage-Aware Speculation Arbitrage：利用优势感知推测实现高效推理 large language model chain-of-thought
8 Decoding the Black Box: Discerning AI Rhetorics About and Through Poetic Prompting 利用诗歌提示模式解码大型语言模型的算法倾向与偏差 large language model
9 LLMs Know More Than Words: A Genre Study with Syntax, Metaphor & Phonetics 提出多语言文类分类数据集，探究LLM对深层语言属性的理解能力 large language model
10 Mitigating Catastrophic Forgetting in Target Language Adaptation of LLMs via Source-Shielded Updates 提出源屏蔽更新方法，缓解LLM目标语言适应中的灾难性遗忘 large language model
11 DAMASHA: Detecting AI in Mixed Adversarial Texts via Segmentation with Human-interpretable Attribution DAMASHA：通过可解释归因分割检测混合对抗文本中的AI生成内容 large language model
12 DaLA: Danish Linguistic Acceptability Evaluation Guided by Real World Errors 提出DaLA，一个基于真实世界丹麦语错误的语言可接受性评估基准。 large language model
13 EtCon: Edit-then-Consolidate for Reliable Knowledge Editing EtCon：编辑-整合范式，提升大语言模型知识编辑的可靠性 large language model
14 SignRoundV2: Closing the Performance Gap in Extremely Low-Bit Post-Training Quantization for LLMs SignRoundV2：弥合LLM极低比特后训练量化中的性能差距 large language model ✅
15 ADAPT: Learning Task Mixtures for Budget-Constrained Instruction Tuning ADAPT：学习任务混合比例，解决预算约束下的指令调优问题。 instruction following
16 AdmTree: Compressing Lengthy Context with Adaptive Semantic Trees AdmTree：提出自适应语义树压缩长文本上下文，提升LLM处理效率。 large language model
17 EvoEdit: Lifelong Free-Text Knowledge Editing through Latent Perturbation Augmentation and Knowledge-driven Parameter Fusion 提出EvoEdit以解决大语言模型知识更新难题 large language model
18 ClusterFusion: Hybrid Clustering with Embedding Guidance and LLM Adaptation 提出ClusterFusion，一种结合嵌入引导和LLM适应的混合聚类框架，提升领域特定文本聚类性能。 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 Semantic Soft Bootstrapping: Long Context Reasoning in LLMs without Reinforcement Learning 提出语义软引导（SSB）方法，无需强化学习即可提升LLM长文本推理能力。 reinforcement learning teacher-student distillation ✅
20 Structured Document Translation via Format Reinforcement Learning 提出FormatRL，通过强化学习优化结构化文档翻译中的格式和语义。 reinforcement learning
21 Nex-N1: Agentic Models Trained via a Unified Ecosystem for Large-Scale Environment Construction Nex-N1：通过统一生态系统训练Agentic模型，用于大规模环境构建 policy learning large language model
22 LangSAT: A Novel Framework Combining NLP and Reinforcement Learning for SAT Solving LangSAT：结合NLP与强化学习的SAT求解新框架 reinforcement learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页