cs.CL（2024-10-30）

📊 共 34 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) (30 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (30 篇)

#题目一句话要点标签🔗⭐
1 Dynamic Strategy Planning for Efficient Question Answering with Large Language Models DyPlan：一种基于大语言模型的动态策略规划方法，提升问答效率与性能。 large language model chain-of-thought
2 Danoliteracy of Generative Large Language Models 提出Danoliteracy基准以评估丹麦语生成模型能力 large language model
3 Smaller Large Language Models Can Do Moral Self-Correction 通过安全对齐微调，小规模语言模型也能实现道德自我修正 large language model
4 On Memorization of Large Language Models in Logical Reasoning 研究表明：大型语言模型在逻辑推理中存在对训练数据的记忆现象 large language model
5 Multi-Agent Large Language Models for Conversational Task-Solving 提出多智能体LLM框架，分析其在会话式任务解决中的优势与挑战。 large language model
6 How Well Do Large Language Models Disambiguate Swedish Words? 评估大型语言模型在瑞典语词义消歧任务中的表现 large language model
7 Constructing Multimodal Datasets from Scratch for Rapid Development of a Japanese Visual Language Model 提出一种从零构建日语多模态数据集的方法，加速日语视觉语言模型开发。 multimodal
8 Linguistics Theory Meets LLM: Code-Switched Text Generation via Equivalence Constrained Large Language Models 提出EZSwitch框架，结合语言学理论与LLM生成高质量Code-Switching文本 large language model
9 Survey of Cultural Awareness in Language Models: Text and Beyond 综述：语言模型中的文化意识——文本及其他 large language model multimodal
10 SciPIP: An LLM-based Scientific Paper Idea Proposer SciPIP：一种基于LLM的科学论文选题推荐框架，提升文献检索与选题生成质量 large language model
11 Comparative Analysis of Demonstration Selection Algorithms for LLM In-Context Learning 对比分析LLM上下文学习的示例选择算法，揭示其效率与有效性差异 large language model ✅
12 Don't Pay Attention, PLANT It: Pretraining Attention via Learning-to-Rank PLANT：通过学习排序预训练注意力机制，提升极端多标签文本分类性能。 large language model ✅
13 Generating Diverse Negations from Affirmative Sentences 提出NegVerse方法，从肯定句生成多样化的否定句，提升语言模型在否定推理方面的能力。 large language model ✅
14 Graph-Augmented Relation Extraction Model with LLMs-Generated Support Document 提出一种图增强的关系抽取模型，利用LLM生成支持文档以提升性能。 large language model
15 ACC-Collab: An Actor-Critic Approach to Multi-Agent LLM Collaboration 提出ACC-Collab，一种基于Actor-Critic的多Agent LLM协作学习框架 large language model
16 Next-Token Prediction Task Assumes Optimal Data Ordering for LLM Training in Proof Generation 针对LLM证明生成，论文提出直观序列数据排序以提升模型训练效果 large language model
17 Eliciting Critical Reasoning in Retrieval-Augmented Language Models via Contrastive Explanations 提出C-RAG框架，通过对比解释提升检索增强语言模型中的批判性推理能力 large language model
18 InjecGuard: Benchmarking and Mitigating Over-defense in Prompt Injection Guardrail Models InjecGuard：通过缓解过防御提升提示注入防御模型的鲁棒性 large language model ✅
19 LLMs as Research Tools: A Large Scale Survey of Researchers' Usage and Perceptions 大规模调研揭示研究人员对LLM作为科研工具的使用情况与看法 large language model
20 Prove Your Point!: Bringing Proof-Enhancement Principles to Argumentative Essay Generation 提出PESA框架，通过论证增强原则提升论证性文章生成的逻辑性和说服力 large language model
21 Neural spell-checker: Beyond words with synthetic data generation 提出基于合成数据生成的神经拼写检查器，显著提升斯洛文尼亚语拼写纠错性能 large language model
22 Collage: Decomposable Rapid Prototyping for Information Extraction on Scientific PDFs Collage：用于科学PDF信息提取的可分解快速原型工具 multimodal
23 Leveraging Language Models and Bandit Algorithms to Drive Adoption of Battery-Electric Vehicles 利用语言模型和Bandit算法个性化推广电动汽车 large language model
24 Can Models Help Us Create Better Models? Evaluating LLMs as Data Scientists 提出FeatEng基准，评估LLM在特征工程代码生成中的能力，助力模型迭代。 large language model
25 Evaluating Cultural and Social Awareness of LLM Web Agents CASA：评估LLM Web Agent文化和社会意识的基准测试 large language model
26 CORAL: Benchmarking Multi-turn Conversational Retrieval-Augmentation Generation 提出CORAL：用于评估多轮对话检索增强生成的大规模基准 large language model
27 Multi-Programming Language Sandbox for LLMs MPLSandbox：为LLM提供多编程语言统一反馈的沙箱环境 large language model
28 Long$^2$RAG: Evaluating Long-Context & Long-Form Retrieval-Augmented Generation with Key Point Recall 提出Long$^2$RAG基准和KPR指标，用于评估长文本检索增强生成模型的性能。 large language model
29 Deep Learning and Machine Learning -- Natural Language Processing: From Theory to Application 探讨深度学习与机器学习在自然语言处理中的应用及挑战 large language model
30 EvoCodeBench: An Evolving Code Generation Benchmark with Domain-Specific Evaluations EvoCodeBench：一个动态演进的代码生成评测基准，具备领域特定评估能力 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🔗⭐
31 Explainable Behavior Cloning: Teaching Large Language Model Agents through Learning by Demonstration 提出EBC-LLMAgent，通过模仿学习提升LLM在移动应用交互中的自主性和可解释性 behavior cloning large language model
32 MDCure: A Scalable Pipeline for Multi-Document Instruction-Following MDCure：一种可扩展的多文档指令跟随流水线，提升LLM在复杂文档集上的处理能力。 PPO instruction following ✅
33 Social Science Meets LLMs: How Reliable Are Large Language Models in Social Simulations? TrustSim：评估LLM在社会模拟中的可靠性，并提出AdaORPO算法提升其性能。 reinforcement learning large language model
34 Less is More: Pre-Training Cross-Lingual Small-Scale Language Models with Cognitively-Plausible Curriculum Learning Strategies 利用认知可信的课程学习策略，预训练跨语言小规模语言模型，提升性能。 curriculum learning

⬅️ 返回 cs.CL 首页 · 🏠 返回主页