cs.CL（2025-12-12）

📊 共 17 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 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (4)

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

#题目一句话要点标签🔗⭐
1 TeleMem: Building Long-Term and Multimodal Memory for Agentic AI TeleMem：构建Agentic AI的长期多模态记忆系统，提升交互性能。 large language model multimodal
2 Minimal Clips, Maximum Salience: Long Video Summarization via Key Moment Extraction 提出基于关键时刻提取的长视频摘要方法，提升视觉信息利用率 large language model multimodal
3 Benchmarking Contextual Understanding for In-Car Conversational Systems 提出基于LLM的评测框架，用于评估车载对话系统中的上下文理解能力。 large language model chain-of-thought
4 VOYAGER: A Training Free Approach for Generating Diverse Datasets using LLMs VOYAGER：一种利用LLM生成多样化数据集的免训练方法 large language model
5 Bounding Hallucinations: Information-Theoretic Guarantees for RAG Systems via Merlin-Arthur Protocols 提出基于Merlin-Arthur协议的RAG训练框架，提升LLM在检索增强生成中的证据依赖性与信息理论保证。 large language model
6 Hold Onto That Thought: Assessing KV Cache Compression On Reasoning 针对长推理任务，评估KV缓存压缩算法对LLM性能的影响 large language model
7 Does Less Hallucination Mean Less Creativity? An Empirical Investigation in LLMs 研究表明，降低LLM幻觉的方法对创造力有不同影响，为科学应用提供指导 large language model
8 CIP: A Plug-and-Play Causal Prompting Framework for Mitigating Hallucinations under Long-Context Noise CIP：一种即插即用的因果提示框架，用于缓解长文本噪声下的幻觉问题 large language model
9 BLASST: Dynamic BLocked Attention Sparsity via Softmax Thresholding BLASST：通过Softmax阈值动态剪枝Attention矩阵，加速长文本LLM推理。 large language model
10 Speculative Decoding Speed-of-Light: Optimal Lower Bounds via Branching Random Walks 提出确定性推测生成算法的运行时间下界 large language model
11 Extending a Parliamentary Corpus with MPs' Tweets: Automatic Annotation and Evaluation Using MultiParTweet 构建多语言议员推文语料库MultiParTweet，融合文本与视觉信息进行情感和主题分析。 multimodal
12 Mistake Notebook Learning: Batch-Clustered Failures for Training-Free Agent Adaptation 提出Mistake Notebook Learning以解决LLM代理自我学习不足问题 large language model ✅
13 AdaSD: Adaptive Speculative Decoding for Efficient Language Model Inference 提出AdaSD自适应推测解码，无需调参提升大语言模型推理效率。 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🔗⭐
14 Direct Confidence Alignment: Aligning Verbalized Confidence with Internal Confidence In Large Language Models 提出直接置信度对齐(DCA)方法，提升大语言模型内部与外部置信度一致性 direct preference optimization large language model
15 When Actions Teach You to Think: Reasoning-Action Synergy via Reinforcement Learning in Conversational Agents 提出基于强化学习的推理-行动协同方法，提升对话Agent的泛化能力。 reinforcement learning large language model
16 Unifying Dynamic Tool Creation and Cross-Task Experience Sharing through Cognitive Memory Architecture SMITH：通过认知记忆架构统一动态工具创建与跨任务经验共享 curriculum learning large language model
17 SUMFORU: An LLM-Based Review Summarization Framework for Personalized Purchase Decision Support SUMFORU：一种基于LLM的评论摘要框架，用于个性化购买决策支持 reinforcement learning distillation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页