cs.CL（2024-09-17）

📊 共 41 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) (36 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱四：生成式动作 (Generative Motion) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (36 篇)

#题目一句话要点标签🔗⭐
1 CoCA: Regaining Safety-awareness of Multimodal Large Language Models with Constitutional Calibration 提出CoCA，通过宪法校准恢复多模态大语言模型对恶意视觉输入的安全性感知。 large language model multimodal
2 NVLM: Open Frontier-Class Multimodal LLMs NVLM 1.0：媲美GPT-4o的前沿多模态大语言模型，提升文本性能并开源 large language model multimodal ✅
3 Chain-of-Thought Prompting for Speech Translation 提出基于思维链提示的语音翻译方法，显著提升Speech-LLM的翻译性能 large language model chain-of-thought
4 Large Language Models are Good Multi-lingual Learners : When LLMs Meet Cross-lingual Prompts 提出MLPrompt多语言提示方法，提升LLM在复杂规则下的推理和理解能力 large language model chain-of-thought
5 Exploring the Trade-Offs: Quantization Methods, Task Difficulty, and Model Size in Large Language Models From Edge to Giant 大规模语言模型量化方法的全面评估：模型大小、任务难度与性能权衡 large language model instruction following
6 Enriching Datasets with Demographics through Large Language Models: What's in a Name? 利用大型语言模型进行人口统计信息推断，提升数据集质量 large language model
7 THaMES: An End-to-End Tool for Hallucination Mitigation and Evaluation in Large Language Models THaMES：用于大规模语言模型幻觉缓解与评估的端到端工具 large language model
8 Task Arithmetic for Language Expansion in Speech Translation 提出增强型任务算术方法，用于语音翻译中的语言扩展，无需重新训练。 large language model foundation model multimodal
9 LOLA -- An Open-Source Massively Multilingual Large Language Model LOLA：一个开源的大规模多语言大型语言模型 large language model
10 The Art of Storytelling: Multi-Agent Generative AI for Dynamic Multimodal Narratives 提出基于多智能体生成式AI的动态多模态叙事教育工具 multimodal
11 Evaluating the Impact of Compression Techniques on Task-Specific Performance of Large Language Models 评估压缩技术对大语言模型任务性能的影响，强调校准数据和评估指标的重要性 large language model
12 Self-Evolutionary Large Language Models through Uncertainty-Enhanced Preference Optimization 提出不确定性增强偏好优化(UPO)，提升LLM自进化性能 large language model
13 Strategic Insights in Human and Large Language Model Tactics at Word Guessing Games 分析人类与大语言模型在猜词游戏中的策略，揭示模型在多语言环境下的挑战。 large language model
14 KVPruner: Structural Pruning for Faster and Memory-Efficient Large Language Models KVPruner：通过结构化剪枝加速并降低大语言模型的内存占用 large language model
15 Enhancing Low-Resource Language and Instruction Following Capabilities of Audio Language Models 提出Typhoon-Audio模型，提升语音语言模型在低资源语言和指令跟随方面的能力 instruction following
16 Enhancing Code-switched Text-to-Speech Synthesis Capability in Large Language Models with only Monolingual Corpora 提出CS-LLM，仅用单语语料提升大语言模型在混合语文本转语音合成中的能力 large language model
17 Investigating Context-Faithfulness in Large Language Models: The Roles of Memory Strength and Evidence Style 研究记忆强度和证据风格对大语言模型上下文忠实度的影响 large language model ✅
18 A Unified Framework to Classify Business Activities into International Standard Industrial Classification through Large Language Models for Circular Economy 利用大型语言模型将商业活动分类到国际标准产业分类，促进循环经济发展。 large language model
19 Adaptive Large Language Models By Layerwise Attention Shortcuts 提出层级注意力捷径，用于自适应大型语言模型计算 large language model
20 Diversify and Conquer: Diversity-Centric Data Selection with Iterative Refinement 提出基于迭代优化的多样性数据选择方法，提升LLM微调效果 large language model instruction following ✅
21 Surveying the MLLM Landscape: A Meta-Review of Current Surveys MLLM综述的元综述：系统性回顾多模态大语言模型评测方法与未来方向 large language model multimodal
22 Less is More: A Simple yet Effective Token Reduction Method for Efficient Multi-modal LLMs 提出TRIM方法，通过CLIP度量进行token缩减，提升多模态LLM效率。 large language model multimodal
23 CREAM: Comparison-Based Reference-Free ELO-Ranked Automatic Evaluation for Meeting Summarization 提出CREAM，一种基于比较和ELO排序的免参考会议摘要自动评估方法 large language model chain-of-thought
24 Towards No-Code Programming of Cobots: Experiments with Code Synthesis by Large Code Models for Conversational Programming 探索基于大型代码模型的对话式编程，实现协作机器人免代码编程 large language model
25 Watch Your Steps: Observable and Modular Chains of Thought 提出程序追踪提示，增强CoT的可观测性和模块化，解决非局部错误问题。 chain-of-thought
26 Small Language Models can Outperform Humans in Short Creative Writing: A Study Comparing SLMs with Humans and LLMs 小语言模型在短篇创意写作中超越人类：SLM与人类及LLM的对比研究 large language model
27 Egalitarian Language Representation in Language Models: It All Begins with Tokenizers 提出GPE，提升语言模型分词器对复杂文字的公平表征 large language model
28 Multi-Document Grounded Multi-Turn Synthetic Dialog Generation 提出一种多文档驱动的多轮合成对话生成技术，提升模型在文档型对话任务上的性能。 chain-of-thought
29 Says Who? Effective Zero-Shot Annotation of Focalization 利用大型语言模型实现叙事焦点零样本标注，性能媲美人工标注。 large language model
30 Reasoning Graph Enhanced Exemplars Retrieval for In-Context Learning 提出RGER：通过推理图增强的范例检索提升上下文学习效果 large language model ✅
31 SC-Phi2: A Fine-tuned Small Language Model for StarCraft II Macromanagement Tasks SC-Phi2：微调的小型语言模型用于星际争霸II的宏观管理任务 large language model
32 Diversity-grounded Channel Prototypical Learning for Out-of-Distribution Intent Detection 提出多样性引导的通道原型学习以解决分布外意图检测问题 large language model
33 DynamicNER: A Dynamic, Multilingual, and Fine-Grained Dataset for LLM-based Named Entity Recognition 提出DynamicNER数据集，用于评估LLM在动态、多语言和细粒度命名实体识别中的能力。 large language model ✅
34 Propulsion: Steering LLM with Tiny Fine-Tuning Propulsion：通过微调缩放LLM特定维度，实现高效任务引导。 large language model
35 Attention-Seeker: Dynamic Self-Attention Scoring for Unsupervised Keyphrase Extraction 提出Attention-Seeker以解决无监督关键短语提取问题 large language model
36 Efficient and Personalized Mobile Health Event Prediction via Small Language Models 利用小型语言模型实现高效且个性化的移动健康事件预测 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🔗⭐
37 Leveraging Distillation Techniques for Document Understanding: A Case Study with FLAN-T5 提出基于蒸馏的文档理解方法，利用FLAN-T5提升文档处理效率。 curriculum learning distillation large language model
38 Bio-Inspired Mamba: Temporal Locality and Bioplausible Learning in Selective State Space Models 提出Bio-Inspired Mamba，融合生物学习原则的在线选择性状态空间模型 Mamba state space model
39 REAL: Response Embedding-based Alignment for LLMs REAL：基于响应嵌入对齐LLM，提升标注效率与模型性能。 RLHF DPO direct preference optimization
40 LLM-as-a-Judge & Reward Model: What They Can and Cannot Do 分析LLM作为评判者和奖励模型的局限性，揭示其在多语言、事实核查和复杂推理上的不足 reinforcement learning large language model

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
41 BAD: Bidirectional Auto-regressive Diffusion for Text-to-Motion Generation 提出双向自回归扩散模型BAD，用于提升文本到动作生成效果 text-to-motion motion generation ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页