cs.CL（2024-12-18）

📊 共 42 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (33 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (9 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (33 篇)

#题目一句话要点标签🔗⭐
1 Typhoon 2: A Family of Open Text and Multimodal Thai Large Language Models Typhoon 2：一系列面向泰语的开源文本和多模态大语言模型 large language model multimodal
2 Fake News Detection: Comparative Evaluation of BERT-like Models and Large Language Models with Generative AI-Annotated Data 利用生成式AI标注数据，对比BERT类模型与大语言模型在假新闻检测中的性能。 large language model
3 Prompting Strategies for Enabling Large Language Models to Infer Causation from Correlation 提出PC-SubQ提示策略，提升大语言模型基于相关性推断因果关系的能力 large language model
4 Memorization Over Reasoning? Exposing and Mitigating Verbatim Memorization in Large Language Models' Character Understanding Evaluation 提出一种缓解LLM在角色理解评估中死记硬背的方法，揭示并减轻逐字记忆的影响。 large language model
5 Hansel: Output Length Controlling Framework for Large Language Models Hansel：一种用于大语言模型输出长度控制的框架 large language model
6 MATCHED: Multimodal Authorship-Attribution To Combat Human Trafficking in Escort-Advertisement Data 提出MATCHED多模态作者归属方法，打击在线护送广告中的人口贩卖 multimodal
7 Physics Reasoner: Knowledge-Augmented Reasoning for Solving Physics Problems with Large Language Models Physics Reasoner：提出知识增强框架，利用大语言模型解决物理问题 large language model
8 Federated Learning and RAG Integration: A Scalable Approach for Medical Large Language Models 提出联邦学习与RAG集成的医学LLM方案，提升性能并保护隐私 large language model
9 Large Language Models for Automated Literature Review: An Evaluation of Reference Generation, Abstract Writing, and Review Composition 评估大型语言模型在文献综述自动化中的能力：参考文献生成、摘要撰写和综述构建 large language model
10 A Statistical and Multi-Perspective Revisiting of the Membership Inference Attack in Large Language Models 针对大型语言模型成员推断攻击，提出一种统计和多视角分析方法，揭示其性能不一致性。 large language model
11 Cracking the Code of Hallucination in LVLMs with Vision-aware Head Divergence 提出视觉感知头差异性度量与强化方法，缓解LVLM中的幻觉问题 large language model multimodal
12 Multi-OphthaLingua: A Multilingual Benchmark for Assessing and Debiasing LLM Ophthalmological QA in LMICs 提出Multi-OphthaLingua多语言眼科QA基准，并设计CLARA方法缓解LLM在低收入国家应用的偏见问题。 large language model chain-of-thought
13 Pipeline Analysis for Developing Instruct LLMs in Low-Resource Languages: A Case Study on Basque 针对低资源语言巴斯克语，提出Instruct LLM开发流程分析与优化方案 large language model instruction following
14 GAMEBoT: Transparent Assessment of LLM Reasoning in Games 提出GAMEBoT以解决LLM推理评估透明性不足问题 large language model chain-of-thought ✅
15 Maximize Your Data's Potential: Enhancing LLM Accuracy with Two-Phase Pretraining 提出两阶段预训练方法，优化数据选择与混合策略，提升LLM准确率。 large language model
16 Digestion Algorithm in Hierarchical Symbolic Forests: A Fast Text Normalization Algorithm and Semantic Parsing Framework for Specific Scenarios and Lightweight Deployment 提出DAHSF算法，结合文本归一化与语义解析，适用于特定场景和轻量级部署。 large language model
17 Channel Merging: Preserving Specialization for Merged Experts 提出通道合并方法，在合并专家模型时保持专业化知识并提升存储效率。 large language model
18 Crabs: Consuming Resource via Auto-generation for LLM-DoS Attack under Black-box Settings 提出AutoDoS，一种黑盒LLM-DoS攻击方法，通过自动生成资源消耗型Prompt实现拒绝服务。 large language model ✅
19 ECG-Byte: A Tokenizer for End-to-End Generative Electrocardiogram Language Modeling 提出ECG-Byte，用于心电图语言建模的端到端Tokenizer，提升训练效率和可解释性。 large language model
20 TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks TheAgentCompany：构建基准测试，评估LLM智能体在真实世界任务中的表现 large language model
21 Towards an optimised evaluation of teachers' discourse: The case of engaging messages 提出一种优化教师话语评估方法，利用大型语言模型识别课堂互动信息。 large language model
22 FarExStance: Explainable Stance Detection for Farsi FarExStance：提出用于波斯语的可解释立场检测数据集与基线模型 large language model
23 A Rose by Any Other Name: LLM-Generated Explanations Are Good Proxies for Human Explanations to Collect Label Distributions on NLI 利用LLM生成解释作为人类解释的替代，用于自然语言推理中标签分布的收集 large language model
24 Domain-adaptative Continual Learning for Low-resource Tasks: Evaluation on Nepali 针对尼泊尔语等低资源任务，提出领域自适应持续学习方法 large language model
25 Meta-Reflection: A Feedback-Free Reflection Learning Framework 提出Meta-Reflection，一种无需反馈的自反思学习框架，提升LLM在电商意图识别等任务中的性能。 large language model
26 Evaluation of LLM Vulnerabilities to Being Misused for Personalized Disinformation Generation 评估大型语言模型在生成个性化虚假信息方面的漏洞，揭示安全过滤器的失效。 large language model
27 PsyDT: Using LLMs to Construct the Digital Twin of Psychological Counselor with Personalized Counseling Style for Psychological Counseling PsyDT：利用LLM构建具有个性化咨询风格的心理咨询师数字孪生 large language model
28 LIFT: Improving Long Context Understanding Through Long Input Fine-Tuning 提出LIFT：通过长输入微调提升长文本理解能力 large language model
29 EvoWiki: Evaluating LLMs on Evolving Knowledge EvoWiki：一个用于评估LLM在演化知识上表现的自动更新数据集 large language model
30 Socio-Culturally Aware Evaluation Framework for LLM-Based Content Moderation 提出一种社会文化敏感的评估框架，用于评估基于LLM的内容审核能力。 large language model
31 EscapeBench: Towards Advancing Creative Intelligence of Language Model Agents EscapeBench：提升语言模型智能体创造性智能的密室逃脱基准 chain-of-thought
32 MetaRuleGPT: Recursive Numerical Reasoning of Language Models Trained with Simple Rules MetaRuleGPT：通过学习简单规则提升语言模型递归数值推理能力 large language model
33 Lightweight Safety Classification Using Pruned Language Models 提出层增强分类(LEC)，利用剪枝语言模型实现轻量级安全内容分类与提示注入检测。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (9 篇)

#题目一句话要点标签🔗⭐
34 A Systematic Examination of Preference Learning through the Lens of Instruction-Following 系统性研究偏好学习数据属性对指令跟随LLM对齐和下游任务性能的影响 preference learning large language model instruction following
35 Inference-Aware Fine-Tuning for Best-of-N Sampling in Large Language Models 提出推理感知微调方法，优化大语言模型Best-of-N采样策略性能 reinforcement learning imitation learning large language model
36 Towards Efficient and Explainable Hate Speech Detection via Model Distillation 提出基于模型蒸馏的高效且可解释的仇恨言论检测方法，提升性能并降低计算成本。 distillation large language model chain-of-thought
37 Context-DPO: Aligning Language Models for Context-Faithfulness 提出Context-DPO，提升大语言模型在检索增强生成中的上下文忠实度 DPO direct preference optimization large language model ✅
38 State Space Models are Strong Text Rerankers 探索状态空间模型在文本重排序中的潜力，性能媲美Transformer。 Mamba SSM state space model
39 Enhancing Knowledge Distillation for LLMs with Response-Priming Prompting 提出响应启动提示的知识蒸馏方法，提升LLM在资源受限环境下的性能 distillation large language model ✅
40 Cross-Lingual Transfer of Debiasing and Detoxification in Multilingual LLMs: An Extensive Investigation 跨语言迁移学习：提升多语言LLM去偏见和去毒性能力的研究 direct preference optimization large language model
41 Curriculum Learning for Cross-Lingual Data-to-Text Generation With Noisy Data 提出基于课程学习的跨语言数据到文本生成方法，提升噪声数据下的生成质量 curriculum learning
42 PLPP: Prompt Learning with Perplexity Is Self-Distillation for Vision-Language Models PLPP：基于困惑度的Prompt学习，实现视觉-语言模型的自蒸馏 distillation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页