cs.CL（2024-10-18）

📊 共 42 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (32 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (32 篇)

#题目一句话要点标签🔗⭐
1 MiCEval: Unveiling Multimodal Chain of Thought's Quality via Image Description and Reasoning Steps 提出MiCEval，通过图像描述和推理步骤评估多模态CoT的质量。 large language model multimodal chain-of-thought ✅
2 Few-Shot Joint Multimodal Entity-Relation Extraction via Knowledge-Enhanced Cross-modal Prompt Model 提出知识增强跨模态Prompt模型，解决少样本多模态实体关系联合抽取问题 large language model multimodal
3 Supervised Chain of Thought 提出监督式思维链（Supervised CoT）方法，提升LLM在复杂推理任务中的性能 large language model chain-of-thought
4 SPRIG: Improving Large Language Model Performance by System Prompt Optimization SPRIG：通过系统提示优化提升大型语言模型性能 large language model
5 Large Language Models Are Overparameterized Text Encoders 通过层剪枝，显著降低大语言模型文本编码的参数冗余，提升推理效率。 large language model
6 DFlow: Diverse Dialogue Flow Simulation with Large Language Models DFlow：利用大语言模型进行多样化对话流模拟，提升任务型对话数据质量。 large language model
7 REEF: Representation Encoding Fingerprints for Large Language Models 提出REEF：一种免训练的大语言模型表征编码指纹方法，用于知识产权保护。 large language model ✅
8 A Lightweight Multi Aspect Controlled Text Generation Solution For Large Language Models 提出一种轻量级数据增强方案，提升大语言模型在多属性控制文本生成任务上的性能。 large language model
9 SylloBio-NLI: Evaluating Large Language Models on Biomedical Syllogistic Reasoning 提出SylloBio-NLI框架，评估大语言模型在生物医学三段论推理中的能力。 large language model
10 Paths-over-Graph: Knowledge Graph Empowered Large Language Model Reasoning 提出Paths-over-Graph (PoG)，通过知识图谱路径增强LLM推理能力，解决幻觉和知识不足问题。 large language model
11 Automated Genre-Aware Article Scoring and Feedback Using Large Language Models 提出一种基于BERT和ChatGPT的自动文章评分与反馈系统，提升文章质量评估的准确性和个性化。 large language model
12 Towards Faithful Natural Language Explanations: A Study Using Activation Patching in Large Language Models 提出因果信实度度量以提升自然语言解释的可信度 large language model ✅
13 CAPE: A Chinese Dataset for Appraisal-based Emotional Generation using Large Language Models CAPE：一个基于认知评估理论的中文情感生成数据集，利用大语言模型提升对话情感表达。 large language model
14 MultiChartQA: Benchmarking Vision-Language Models on Multi-Chart Problems 提出MultiChartQA基准，评估视觉语言模型在多图表推理中的能力 large language model multimodal ✅
15 MoDification: Mixture of Depths Made Easy MoDification：一种简易的深度混合方法，提升长文本LLM效率 large language model
16 LabSafety Bench: Benchmarking LLMs on Safety Issues in Scientific Labs 提出LabSafety Bench，评估LLM在科学实验室安全问题上的表现 large language model
17 SudoLM: Learning Access Control of Parametric Knowledge with Authorization Alignment 提出SudoLM，通过授权对齐实现LLM参数知识的访问控制。 large language model
18 Distance between Relevant Information Pieces Causes Bias in Long-Context LLMs LongPiBench：揭示长文本LLM中相关信息间距引起的偏差问题 large language model
19 Teaching Models to Balance Resisting and Accepting Persuasion 提出Persuasion-Training (PBT)方法，提升LLM在对抗性说服中的抵抗力与接受有益说服的能力。 large language model
20 Tell me what I need to know: Exploring LLM-based (Personalized) Abstractive Multi-Source Meeting Summarization 提出基于LLM的三阶段方法，实现个性化多源会议摘要生成 large language model
21 Rationale Behind Essay Scores: Enhancing S-LLM's Multi-Trait Essay Scoring with Rationale Generated by LLMs 提出基于理由的多维度作文评分方法RMTS，利用LLM提升S-LLM的评分性能。 large language model ✅
22 Are AI Detectors Good Enough? A Survey on Quality of Datasets With Machine-Generated Texts 针对AI生成文本检测数据集质量评估，提出系统性评测方法以提升检测器泛化能力。 large language model ✅
23 Combining Entropy and Matrix Nuclear Norm for Enhanced Evaluation of Language Models 结合熵与矩阵核范数，提出一种增强型语言模型评估方法 large language model
24 A Systematic Study of Cross-Layer KV Sharing for Efficient LLM Inference 系统性研究跨层KV共享，提升LLM推理效率 large language model
25 How Do Multilingual Language Models Remember Facts? 揭示多语言LLM事实记忆机制：语言依赖与独立性分析 large language model
26 Critical Questions Generation: Motivation and Challenges 提出关键问题生成任务，利用LLM质疑论证，缓解其知识过时和幻觉问题。 large language model
27 EcomEdit: An Automated E-commerce Knowledge Editing Framework for Enhanced Product and Purchase Intention Understanding 提出ECOMEDIT，一个电商知识编辑框架，提升LLM对商品和购买意图的理解。 large language model
28 Good Parenting is all you need -- Multi-agentic LLM Hallucination Mitigation 多智能体LLM协同通过“好家长”式反馈缓解幻觉问题 large language model
29 Beyond Binary: Towards Fine-Grained LLM-Generated Text Detection via Role Recognition and Involvement Measurement 提出LLMDetect基准，通过角色识别和参与度测量实现细粒度的LLM生成文本检测。 large language model
30 Towards Robust Knowledge Representations in Multilingual LLMs for Equivalence and Inheritance based Consistent Reasoning 提出组合表示方法，提升多语言LLM在等价性和继承性推理任务中的一致性。 large language model
31 XForecast: Evaluating Natural Language Explanations for Time Series Forecasting XForecast提出基于可模拟性的指标，评估时间序列预测自然语言解释的质量。 large language model
32 SRAP-Agent: Simulating and Optimizing Scarce Resource Allocation Policy with LLM-based Agent SRAP-Agent：利用LLM模拟和优化稀缺资源分配策略 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
33 MetaAlign: Align Large Language Models with Diverse Preferences during Inference Time MetaAlign：推理时对齐大语言模型与多样化偏好 reinforcement learning RLHF DPO
34 To Trust or Not to Trust? Enhancing Large Language Models' Situated Faithfulness to External Contexts 提升大语言模型情境忠实度：基于置信度推理解决外部知识冲突 DPO direct preference optimization large language model
35 Unveiling Large Language Models Generated Texts: A Multi-Level Fine-Grained Detection Framework 提出多层次细粒度检测框架MFD，有效识别大型语言模型生成的文本 MAE contrastive learning large language model
36 LoGU: Long-form Generation with Uncertainty Expressions 提出LoGU框架，通过不确定性表达提升长文本生成的事实准确性并减少幻觉。 DPO direct preference optimization large language model
37 DiSCo: LLM Knowledge Distillation for Efficient Sparse Retrieval in Conversational Search DiSCo：面向会话搜索，通过LLM知识蒸馏实现高效稀疏检索 representation learning distillation large language model
38 TreeBoN: Enhancing Inference-Time Alignment with Speculative Tree-Search and Best-of-N Sampling TreeBoN：通过推断时投机树搜索和Best-of-N采样增强大语言模型对齐 RLHF DPO direct preference optimization
39 Unlearning Backdoor Attacks for LLMs with Weak-to-Strong Knowledge Distillation 提出W2SDefense，利用弱到强知识蒸馏消除LLM中的后门攻击 distillation large language model
40 Montessori-Instruct: Generate Influential Training Data Tailored for Student Learning Montessori-Instruct：通过定制化数据合成提升学生语言模型的学习效果 DPO direct preference optimization large language model ✅
41 Controllable Discovery of Intents: Incremental Deep Clustering Using Semi-Supervised Contrastive Learning 提出CDI框架，通过半监督对比学习实现可控的意图发现 contrastive learning
42 GenEOL: Harnessing the Generative Power of LLMs for Training-Free Sentence Embeddings GenEOL：利用LLM的生成能力，实现免训练的句子嵌入 contrastive learning large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页