cs.CL（2024-10-11）

📊 共 41 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) (36 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗2) 支柱一：机器人控制 (Robot Control) (2 🔗1)

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

#题目一句话要点标签🔗⭐
1 M3Hop-CoT: Misogynous Meme Identification with Multimodal Multi-hop Chain-of-Thought 提出M3Hop-CoT框架，利用多模态多跳思维链识别仇恨女性的Meme。 large language model multimodal chain-of-thought
2 Large Language Models for Medical OSCE Assessment: A Novel Approach to Transcript Analysis 利用大型语言模型进行医学OSCE评估，实现病史总结能力自动评分 large language model chain-of-thought
3 More than Memes: A Multimodal Topic Modeling Approach to Conspiracy Theories on Telegram 提出基于BERTopic和CLIP的多模态主题建模方法，分析Telegram阴谋论内容。 multimodal
4 LLMD: A Large Language Model for Interpreting Longitudinal Medical Records LLMD：用于解读纵向医疗记录的大语言模型 large language model
5 Enterprise Benchmarks for Large Language Model Evaluation 提出企业级LLM评测基准，涵盖金融、法律、网络安全等领域 large language model
6 NoVo: Norm Voting off Hallucinations with Attention Heads in Large Language Models 提出NoVo，利用注意力头范数投票显著提升大语言模型的事实准确性 large language model
7 Developing a Pragmatic Benchmark for Assessing Korean Legal Language Understanding in Large Language Models 提出KBL：用于评估大型语言模型韩语法律语言理解能力的实用基准 large language model
8 Humanity in AI: Detecting the Personality of Large Language Models 结合文本挖掘与问卷调查，提升大语言模型人格检测的可靠性。 large language model
9 Exploring the Role of Reasoning Structures for Constructing Proofs in Multi-Step Natural Language Reasoning with Large Language Models 探索推理结构在LLM多步自然语言推理证明构建中的作用 large language model
10 oRetrieval Augmented Generation for 10 Large Language Models and its Generalizability in Assessing Medical Fitness 利用检索增强生成技术（RAG）提升大型语言模型在医疗健康领域的适应性，尤其是在术前评估方面。 large language model
11 Sui Generis: Large Language Models for Authorship Attribution and Verification in Latin 利用大型语言模型解决拉丁语文本的作者身份归属与验证问题 large language model
12 Fine-Tuning In-House Large Language Models to Infer Differential Diagnosis from Radiology Reports 提出一种基于自研LLM的放射报告差异诊断推断微调方案，性能媲美GPT-4。 large language model
13 Hypothesis-only Biases in Large Language Model-Elicited Natural Language Inference 揭示大语言模型生成NLI数据中的假设偏差，强调数据质量对模型评估的重要性 large language model
14 Measuring the Inconsistency of Large Language Models in Preferential Ranking 评估大语言模型在偏好排序中的一致性问题，揭示其内在缺陷 large language model
15 A social context-aware graph-based multimodal attentive learning framework for disaster content classification during emergencies: a benchmark dataset and method 提出CrisisSpot框架，利用社交上下文感知图神经网络进行紧急事件中灾害内容分类。 multimodal
16 SocialGaze: Improving the Integration of Human Social Norms in Large Language Models 提出SocialGaze框架，提升大语言模型对人类社会规范的理解与对齐 large language model
17 Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuning 提出语义知识调优(SK-Tuning)，高效微调大语言模型，提升文本理解和分类性能。 large language model
18 Audio Description Generation in the Era of LLMs and VLMs: A Review of Transferable Generative AI Technologies 综述性论文：探讨大语言模型和视觉语言模型在自动语音描述生成中的应用 large language model
19 Nudging: Inference-time Alignment of LLMs via Guided Decoding 提出NUDGING：一种基于引导解码的LLM推理期对齐方法 large language model ✅
20 Science is Exploration: Computational Frontiers for Conceptual Metaphor Theory 利用大型语言模型探索概念隐喻理论的计算前沿 large language model
21 Towards Multilingual LLM Evaluation for European Languages 提出面向欧洲语言的多语言LLM评估框架，解决跨语种性能评估难题。 large language model
22 QEFT: Quantization for Efficient Fine-Tuning of LLMs QEFT：一种高效微调LLM的量化方法，兼顾推理效率与模型质量 large language model ✅
23 The Same But Different: Structural Similarities and Differences in Multilingual Language Modeling 利用机制可解释性探究多语言模型中语言结构的处理方式 large language model
24 Enhancing Long Context Performance in LLMs Through Inner Loop Query Mechanism 提出基于内循环查询机制的ILM-TR模型，提升LLM在长文本环境下的性能 large language model
25 SimpleStrat: Diversifying Language Model Generation with Stratification SimpleStrat：通过分层抽样提升语言模型生成的多样性 large language model
26 Controllable Safety Alignment: Inference-Time Adaptation to Diverse Safety Requirements 提出CoSA框架，通过推理时调整安全配置，实现LLM对多样化安全需求的可控对齐 large language model
27 A Benchmark for Cross-Domain Argumentative Stance Classification on Social Media 提出一种基于平台规则和LLM的多领域论证立场分类基准 large language model
28 StructRAG: Boosting Knowledge Intensive Reasoning of LLMs via Inference-time Hybrid Information Structurization StructRAG：通过推理时混合信息结构化增强LLM的知识密集型推理能力 large language model
29 Scaling Laws for Predicting Downstream Performance in LLMs 提出FLP和FLP-M方法，利用预训练损失预测LLM下游任务性能，降低计算成本。 large language model
30 Hybrid Training Approaches for LLMs: Leveraging Real and Synthetic Data to Enhance Model Performance in Domain-Specific Applications 提出混合训练方法，利用真实和合成数据提升LLM在领域特定应用中的性能 large language model
31 The Impact of Visual Information in Chinese Characters: Evaluating Large Models' Ability to Recognize and Utilize Radicals 评估大模型对汉字视觉信息的理解：利用部首提升中文处理任务 large language model
32 RoRA-VLM: Robust Retrieval-Augmented Vision Language Models 提出RoRA-VLM，增强视觉语言模型在知识密集型任务中的检索能力和鲁棒性 multimodal
33 Which Demographics do LLMs Default to During Annotation? 研究LLM在无人口统计信息条件下的默认标注倾向，揭示其内在偏见 large language model
34 Data Processing for the OpenGPT-X Model Family OpenGPT-X项目：构建大规模多语种LLM的数据处理流程 large language model
35 AMPO: Automatic Multi-Branched Prompt Optimization 提出AMPO，一种自动多分支提示优化方法，提升LLM在复杂任务中的性能。 large language model
36 StraGo: Harnessing Strategic Guidance for Prompt Optimization StraGo：利用策略指导优化提示，解决提示漂移问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
37 SuperCorrect: Advancing Small LLM Reasoning with Thought Template Distillation and Self-Correction SuperCorrect：通过思想模板蒸馏和自校正提升小LLM的推理能力 DPO direct preference optimization distillation ✅
38 Mentor-KD: Making Small Language Models Better Multi-step Reasoners 提出Mentor-KD，通过中间导师模型提升小语言模型的多步推理能力 distillation large language model chain-of-thought
39 Language Imbalance Driven Rewarding for Multilingual Self-improving 提出语言不平衡驱动的奖励机制，用于多语言大模型的自提升。 DPO large language model instruction following ✅

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
40 AttnGCG: Enhancing Jailbreaking Attacks on LLMs with Attention Manipulation AttnGCG：通过注意力操纵增强大语言模型的越狱攻击 manipulation large language model ✅
41 Unraveling and Mitigating Safety Alignment Degradation of Vision-Language Models 提出跨模态表征操控(CMRM)方法，缓解视觉语言模型中的安全性对齐退化问题 manipulation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页