cs.CL（2025-10-21）

📊 共 36 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (29 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗2) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (29 篇)

#题目一句话要点标签🔗⭐
1 ECG-LLM -- training and evaluation of domain-specific large language models for electrocardiography ECG-LLM：心电图领域专用大语言模型的训练与评估 large language model multimodal
2 Text or Pixels? It Takes Half: On the Token Efficiency of Visual Text Inputs in Multimodal LLMs 提出文本图像化压缩方法，在多模态LLM中降低Token使用量并保持性能 large language model multimodal
3 Chain-of-Conceptual-Thought Elicits Daily Conversation in Large Language Models 提出概念链式思考(CoCT)提示方法，提升大语言模型在日常对话中的表现 large language model chain-of-thought
4 A Graph Signal Processing Framework for Hallucination Detection in Large Language Models 提出基于图信号处理的框架，用于检测大型语言模型中的幻觉现象 large language model
5 Large language models for folktale type automation based on motifs: Cinderella case study 利用大型语言模型和主题自动化分析灰姑娘故事变体 large language model
6 Misinformation Detection using Large Language Models with Explainability 提出一种可解释的轻量级PLM框架，用于高效且可信地检测虚假信息。 large language model
7 From Memorization to Generalization: Fine-Tuning Large Language Models for Biomedical Term-to-Identifier Normalization 微调大型语言模型用于生物医学术语-标识符归一化，揭示泛化能力差异 large language model
8 Identity-Aware Large Language Models require Cultural Reasoning 提出文化推理能力，解决大语言模型中身份感知不足的问题 large language model
9 BrailleLLM: Braille Instruction Tuning with Large Language Models for Braille Domain Tasks BrailleLLM：通过大型语言模型和盲文指令调优，解决盲文领域任务 large language model
10 Are they lovers or friends? Evaluating LLMs' Social Reasoning in English and Korean Dialogues SCRIPTS：评估LLM在英韩对话中进行社会推理能力的数据集与分析 large language model chain-of-thought
11 Grounding or Guessing? Visual Signals for Detecting Hallucinations in Sign Language Translation 提出基于视觉信号的可靠性度量，用于检测手语翻译中的幻觉问题。 multimodal visual grounding
12 LightMem: Lightweight and Efficient Memory-Augmented Generation LightMem：一种轻量高效的记忆增强生成模型，提升LLM在动态环境中的性能。 large language model ✅
13 MTraining: Distributed Dynamic Sparse Attention for Efficient Ultra-Long Context Training MTraining：分布式动态稀疏注意力加速超长上下文LLM训练 large language model ✅
14 Evaluating LLM Story Generation through Large-scale Network Analysis of Social Structures 提出基于社交结构网络分析的大规模LLM故事生成评估方法 large language model
15 Bayesian Low-Rank Factorization for Robust Model Adaptation 提出基于贝叶斯低秩分解的适配器，用于稳健地适应语音基础模型，解决代码切换场景下的过拟合问题。 foundation model
16 DuoLens: A Framework for Robust Detection of Machine-Generated Multilingual Text and Code 提出DuoLens框架以解决机器生成多语言文本和代码检测问题 large language model
17 How Do LLMs Use Their Depth? 提出“猜测-精炼”框架，揭示LLM逐层预测的动态过程与深度利用方式 large language model
18 UNO-Bench: A Unified Benchmark for Exploring the Compositional Law Between Uni-modal and Omni-modal in Omni Models 提出UNO-Bench，用于统一评估全模态模型中单模态与多模态的组合规律。 multimodal
19 Context-aware Fairness Evaluation and Mitigation in LLMs 提出上下文感知动态剪枝框架，用于大语言模型中的公平性评估与缓解。 large language model
20 Learning from the Best, Differently: A Diversity-Driven Rethinking on Data Selection 提出ODiS算法，通过正交解耦数据质量评估维度，提升LLM预训练数据选择的多样性与质量。 large language model
21 Improving Topic Modeling of Social Media Short Texts with Rephrasing: A Case Study of COVID-19 Related Tweets 提出TM-Rephrase框架，利用LLM重述社交媒体文本，提升主题建模效果。 large language model
22 DelvePO: Direction-Guided Self-Evolving Framework for Flexible Prompt Optimization DelvePO：面向灵活提示优化的方向引导自进化框架 large language model
23 Contrastive Decoding Mitigates Score Range Bias in LLM-as-a-Judge 对比解码缓解LLM评分中的范围偏差，提升LLM作为裁判的可靠性 large language model
24 When Can We Trust LLMs in Mental Health? Large-Scale Benchmarks for Reliable LLM Evaluation 提出MentalBench和MentalAlign，用于可靠评估LLM在心理健康领域的应用 large language model ✅
25 KAT-Coder Technical Report KAT-Coder：通过多阶段训练提升LLM在交互式软件开发中的自主编码能力 large language model ✅
26 CEFR-Annotated WordNet: LLM-Based Proficiency-Guided Semantic Database for Language Learning 提出基于LLM的CEFR标注WordNet，用于提升语言学习效果 large language model
27 KoSimpleQA: A Korean Factuality Benchmark with an Analysis of Reasoning LLMs 提出KoSimpleQA基准，用于评估LLM在韩语事实性知识问答中的表现 large language model
28 Combining Distantly Supervised Models with In Context Learning for Monolingual and Cross-Lingual Relation Extraction 提出HYDRE框架，结合远程监督模型与上下文学习，提升单语和跨语关系抽取效果。 large language model
29 Position: LLM Watermarking Should Align Stakeholders' Incentives for Practical Adoption 提出激励对齐的LLM水印方案，促进实际应用落地 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🔗⭐
30 Fine-Tuned Thoughts: Leveraging Chain-of-Thought Reasoning for Industrial Asset Health Monitoring 提出基于CoT蒸馏的知识迁移框架，提升SLM在工业资产健康监测中的推理能力。 distillation large language model chain-of-thought ✅
31 Towards Faithful and Controllable Personalization via Critique-Post-Edit Reinforcement Learning 提出Critique-Post-Edit强化学习框架，实现更忠实和可控的LLM个性化。 reinforcement learning PPO RLHF
32 MENTOR: A Reinforcement Learning Framework for Enabling Tool Use in Small Models via Teacher-Optimized Rewards MENTOR：一种通过教师优化奖励在小模型中实现工具使用的强化学习框架 reinforcement learning distillation large language model
33 Verifiable Accuracy and Abstention Rewards in Curriculum RL to Alleviate Lost-in-Conversation 提出RLAAR框架，通过可验证奖励的课程强化学习缓解多轮对话中的信息丢失问题。 reinforcement learning large language model instruction following
34 Every Step Evolves: Scaling Reinforcement Learning for Trillion-Scale Thinking Model 提出Ring-1T：一个具有万亿参数的开源思维模型，解决训练和推理不一致等挑战。 reinforcement learning IMoS
35 WebSeer: Training Deeper Search Agents through Reinforcement Learning with Self-Reflection WebSeer：通过自反思强化学习训练更深度的搜索Agent reinforcement learning ✅

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

#题目一句话要点标签🔗⭐
36 Engagement Undermines Safety: How Stereotypes and Toxicity Shape Humor in Language Models 评估幽默生成中的刻板印象与毒性对安全性的影响 HuMoR large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页