cs.CL（2024-08-16）

📊 共 26 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (22 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (2) 支柱三：空间感知与语义 (Perception & Semantics) (1) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (22 篇)

#题目一句话要点标签🔗⭐
1 A Survey on Benchmarks of Multimodal Large Language Models 多模态大语言模型评测基准综述：全面评估与未来方向 large language model multimodal ✅
2 Math-PUMA: Progressive Upward Multimodal Alignment to Enhance Mathematical Reasoning Math-PUMA：通过渐进式向上多模态对齐增强数学推理能力 large language model multimodal ✅
3 When Prompting Fails to Sway: Inertia in Moral and Value Judgments of Large Language Models 揭示大型语言模型在道德和价值判断中存在的惯性，即使通过prompt干预。 large language model
4 PEDAL: Enhancing Greedy Decoding with Large Language Models using Diverse Exemplars PEDAL：利用多样化范例增强大语言模型贪婪解码，提升文本生成性能 large language model
5 PsychoLex: Unveiling the Psychological Mind of Large Language Models PsychoLex：构建并评估面向心理学任务的波斯语和英语大型语言模型 large language model
6 Large Language Models Might Not Care What You Are Saying: Prompt Format Beats Descriptions 提出集成提示框架，发现大语言模型对提示描述内容不敏感，提示格式更重要。 large language model
7 Turning Trash into Treasure: Accelerating Inference of Large Language Models with Token Recycling 提出Token Recycling，加速大语言模型推理，无需额外训练。 large language model
8 MIA-Tuner: Adapting Large Language Models as Pre-training Text Detector MIA-Tuner：利用指令调优大语言模型作为预训练文本检测器 large language model
9 Chain of Thought Still Thinks Fast: APriCoT Helps with Thinking Slow 提出APriCoT方法，缓解语言模型在MMLU任务中的偏差，提升推理的稳健性。 chain-of-thought
10 Using large language models to estimate features of multi-word expressions: Concreteness, valence, arousal 利用大型语言模型评估多词表达的具象性、效价和唤醒度 large language model
11 Collaborative Cross-modal Fusion with Large Language Model for Recommendation 提出CCF-LLM框架，通过协同跨模态融合增强LLM在推荐系统中的性能。 large language model
12 SelectLLM: Query-Aware Efficient Selection Algorithm for Large Language Models SelectLLM：一种查询感知的LLM高效选择算法，提升推理效率。 large language model
13 MuRAR: A Simple and Effective Multimodal Retrieval and Answer Refinement Framework for Multimodal Question Answering MuRAR：一个简单高效的多模态检索与答案优化框架，用于多模态问答 multimodal
14 Med-PMC: Medical Personalized Multi-modal Consultation with a Proactive Ask-First-Observe-Next Paradigm 提出Med-PMC评估框架，用于评估多模态大语言模型在医疗个性化多模态咨询中的临床能力。 large language model multimodal ✅
15 LLMs Are Biased Towards Output Formats! Systematically Evaluating and Mitigating Output Format Bias of LLMs 系统评估并缓解LLM的输出格式偏差，提升模型在不同格式下的性能一致性。 large language model instruction following
16 Ex3: Automatic Novel Writing by Extracting, Excelsior and Expanding Ex3：通过抽取、精进与扩展实现自动小说创作 large language model instruction following
17 FLEXTAF: Enhancing Table Reasoning with Flexible Tabular Formats FLEXTAF：通过灵活表格格式增强表格推理能力 large language model
18 EmoDynamiX: Emotional Support Dialogue Strategy Prediction by Modelling MiXed Emotions and Discourse Dynamics EmoDynamiX：通过混合情绪和对话动态建模预测情感支持对话策略 large language model
19 DAC: Decomposed Automation Correction for Text-to-SQL 提出分解自动化纠错(DAC)方法，提升Text-to-SQL任务中LLM的SQL生成质量。 large language model
20 Lower Layers Matter: Alleviating Hallucination via Multi-Layer Fusion Contrastive Decoding with Truthfulness Refocused 提出LOL框架，通过多层融合对比解码和真值重聚焦缓解大语言模型的幻觉问题 large language model
21 The Fellowship of the LLMs: Multi-Model Workflows for Synthetic Preference Optimization Dataset Generation 提出基于多模型工作流的合成偏好优化数据集生成方法，提升数据集构建效率。 large language model
22 CommunityKG-RAG: Leveraging Community Structures in Knowledge Graphs for Advanced Retrieval-Augmented Generation in Fact-Checking 提出CommunityKG-RAG，利用知识图谱社区结构增强事实核查中的RAG性能 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#题目一句话要点标签🔗⭐
23 ChatZero:Zero-shot Cross-Lingual Dialogue Generation via Pseudo-Target Language ChatZero：通过伪目标语言实现零样本跨语言对话生成 contrastive learning large language model
24 Where is the signal in tokenization space? 探索tokenization空间中的信号：通过非规范token化提升LLM性能 state space model large language model

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
25 From Lazy to Prolific: Tackling Missing Labels in Open Vocabulary Extreme Classification by Positive-Unlabeled Sequence Learning 提出正负序列学习以解决开放词汇极端分类中的缺失标签问题 open-vocabulary open vocabulary

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
26 See What LLMs Cannot Answer: A Self-Challenge Framework for Uncovering LLM Weaknesses 提出自挑战框架，揭示大语言模型自身弱点并构建评测基准。 manipulation large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页