cs.CL（2024-12-17）

📊 共 30 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (25 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (25 篇)

#题目一句话要点标签🔗⭐
1 Question: How do Large Language Models perform on the Question Answering tasks? Answer: 对比研究：大型语言模型在问答任务中的表现及单次推理提示优化 large language model instruction following
2 DateLogicQA: Benchmarking Temporal Biases in Large Language Models 提出DateLogicQA基准测试，用于评估大型语言模型中的时间推理偏差。 large language model
3 Refining Answer Distributions for Improved Large Language Model Reasoning 提出精炼答案分布法，提升大语言模型推理能力 large language model
4 Compressed Chain of Thought: Efficient Reasoning Through Dense Representations 提出压缩思维链（CCoT），通过稠密表示提升语言模型推理效率。 chain-of-thought
5 Algorithmic Fidelity of Large Language Models in Generating Synthetic German Public Opinions: A Case Study 研究大型语言模型在生成德国公众意见中的算法忠实性 large language model
6 CLASP: Contrastive Language-Speech Pretraining for Multilingual Multimodal Information Retrieval CLASP：用于多语言多模态信息检索的对比语言-语音预训练 multimodal
7 RCLMuFN: Relational Context Learning and Multiplex Fusion Network for Multimodal Sarcasm Detection 提出RCLMuFN模型，通过关系上下文学习和多路复用融合提升多模态讽刺检测性能。 multimodal
8 SnakModel: Lessons Learned from Training an Open Danish Large Language Model SnakModel：基于Llama2-7B的丹麦语大语言模型训练与优化实践 large language model
9 DSGram: Dynamic Weighting Sub-Metrics for Grammatical Error Correction in the Era of Large Language Models 提出DSGram框架，利用动态权重子指标提升大语言模型时代语法纠错评估的有效性 large language model
10 RAG-Star: Enhancing Deliberative Reasoning with Retrieval Augmented Verification and Refinement RAG-Star：通过检索增强的验证与改进提升LLM的审慎推理能力 large language model chain-of-thought
11 An Automated Explainable Educational Assessment System Built on LLMs AERA Chat：基于LLM的自动化、可解释教育评估系统 large language model
12 Extending LLMs to New Languages: A Case Study of Llama and Persian Adaptation 提出多阶段参数高效微调方法，扩展Llama模型至波斯语 large language model
13 Training Dynamics of a 1.7B LLaMa Model: A Data-Efficient Approach 训练17亿参数LLaMa模型：一种数据高效的方法 large language model ✅
14 DnDScore: Decontextualization and Decomposition for Factuality Verification in Long-Form Text Generation 提出DnDScore，通过解构和去语境化提升长文本生成的事实性验证效果 large language model
15 Memory-Augmented Agent Training for Business Document Understanding 提出Matrix框架，通过记忆增强Agent训练提升LLM在商业文档理解中的性能 large language model
16 AI PERSONA: Towards Life-long Personalization of LLMs 提出AI Persona框架，实现大语言模型（LLM）的终身个性化 large language model
17 AIR-Bench: Automated Heterogeneous Information Retrieval Benchmark 提出AIR-Bench：自动化异构信息检索评测基准，解决新兴领域评测难题 large language model ✅
18 NAVCON: A Cognitively Inspired and Linguistically Grounded Corpus for Vision and Language Navigation 提出NAVCON：一个认知启发且语言对齐的视觉语言导航语料库 VLN
19 Unlocking LLMs: Addressing Scarce Data and Bias Challenges in Mental Health 提出IC-AnnoMI数据集，解决心理健康领域LLM数据稀缺和偏差问题 large language model
20 Adaptations of AI models for querying the LandMatrix database in natural language 利用AI模型适配LandMatrix数据库，实现自然语言查询 large language model ✅
21 Truthful Text Sanitization Guided by Inference Attacks 提出基于推理攻击指导的文本泛化脱敏方法，平衡隐私保护与效用 large language model
22 Benchmarking and Understanding Compositional Relational Reasoning of LLMs 提出GAR基准测试，用于评估和理解LLM的组合关系推理能力 large language model ✅
23 More Tokens, Lower Precision: Towards the Optimal Token-Precision Trade-off in KV Cache Compression 提出量化剪枝方法，优化KV缓存压缩中的Token-精度权衡，提升长文本LLM性能 large language model ✅
24 Trigger$^3$: Refining Query Correction via Adaptive Model Selector 提出Trigger$^3$，通过自适应模型选择优化查询纠错。 large language model
25 Exploring Cross-lingual Latent Transplantation: Mutual Opportunities and Open Challenges 提出跨语言隐空间迁移框架，提升LLM多语言能力和文化适应性 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🔗⭐
26 Preference-Oriented Supervised Fine-Tuning: Favoring Target Model Over Aligned Large Language Models 提出偏好导向的监督微调方法以提升模型性能 DPO large language model
27 Expansion Span: Combining Fading Memory and Retrieval in Hybrid State Space Models 提出Span-Expanded Attention，扩展混合状态空间模型的记忆范围，提升长序列建模能力。 SSM state space model
28 In-Context Learning Distillation for Efficient Few-Shot Fine-Tuning 提出上下文学习蒸馏方法，实现高效小样本微调并显著压缩模型规模。 distillation
29 Detecting Emotional Incongruity of Sarcasm by Commonsense Reasoning 提出基于常识推理的情感不一致性检测框架EICR，用于识别反讽语句 contrastive 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🔗⭐
30 Revealing the impact of synthetic native samples and multi-tasking strategies in Hindi-English code-mixed humour and sarcasm detection 提出融合原生样本与多任务学习策略，提升印地语-英语混合语幽默与讽刺检测性能 HuMoR

⬅️ 返回 cs.CL 首页 · 🏠 返回主页