cs.CL（2024-05-28）

📊 共 43 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) (33 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗2) 支柱七：动作重定向 (Motion Retargeting) (1) 支柱一：机器人控制 (Robot Control) (1)

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

#题目一句话要点标签🔗⭐
1 Faithful Logical Reasoning via Symbolic Chain-of-Thought 提出SymbCoT框架，结合符号逻辑推理增强大语言模型的逻辑推理能力 large language model chain-of-thought ✅
2 Understanding Intrinsic Socioeconomic Biases in Large Language Models 揭示大语言模型中固有的社会经济偏见，关注交叉性影响。 large language model
3 Exploring Context Window of Large Language Models via Decomposed Positional Vectors 通过分解位置向量探索大语言模型上下文窗口，并提出无训练扩展方法。 large language model
4 Personalized Steering of Large Language Models: Versatile Steering Vectors Through Bi-directional Preference Optimization 提出双向偏好优化方法，生成可控的LLM个性化引导向量。 large language model
5 ConSiDERS-The-Human Evaluation Framework: Rethinking Human Evaluation for Generative Large Language Models 提出ConSiDERS框架，用于改进生成式大语言模型的人工评估方法 large language model
6 An Empirical Analysis on Large Language Models in Debate Evaluation 研究表明大型语言模型在辩论评估中表现优异，但存在多种偏见 large language model
7 Active Use of Latent Constituency Representation in both Humans and Large Language Models 通过单样本学习任务，揭示人类和大型语言模型中潜在的成分句法结构 large language model
8 Edinburgh Clinical NLP at MEDIQA-CORR 2024: Guiding Large Language Models with Hints 利用提示引导大型语言模型，解决临床笔记中的医疗错误识别与纠正问题。 large language model
9 Tool Learning with Large Language Models: A Survey 综述：大型语言模型工具学习，提升复杂问题解决能力 large language model ✅
10 Conv-CoA: Improving Open-domain Question Answering in Large Language Models via Conversational Chain-of-Action 提出Conv-CoA框架，通过对话式行动链提升大语言模型在开放域问答中的表现。 large language model
11 IAPT: Instruction-Aware Prompt Tuning for Large Language Models 提出指令感知Prompt Tuning（IAPT），仅用四个soft token实现高效LLM微调。 large language model
12 TimeChara: Evaluating Point-in-Time Character Hallucination of Role-Playing Large Language Models 提出TimeChara基准，评估角色扮演大语言模型的时间点幻觉问题，并提出Narrative-Experts方法缓解。 large language model
13 C$^{3}$Bench: A Comprehensive Classical Chinese Understanding Benchmark for Large Language Models 提出C³Bench，用于全面评估大语言模型在古文理解方面的能力 large language model ✅
14 Arithmetic Reasoning with LLM: Prolog Generation & Permutation 提出基于Prolog生成的算术推理方法，提升LLM在数学问题上的表现。 large language model chain-of-thought
15 MMCTAgent: Multi-modal Critical Thinking Agent Framework for Complex Visual Reasoning 提出MMCTAgent，用于复杂视觉推理的多模态批判性思维Agent框架 large language model
16 Knowledge Circuits in Pretrained Transformers 揭示Transformer中知识存储方式：提出知识回路以理解模型行为 large language model ✅
17 Don't Forget to Connect! Improving RAG with Graph-based Reranking 提出G-RAG：一种基于图神经网络的RAG重排序方法，提升文档连接和语义理解能力 large language model
18 PromptWizard: Task-Aware Prompt Optimization Framework PromptWizard：面向任务的自适应提示优化框架，提升LLM性能。 large language model
19 The Impossibility of Fair LLMs 论证通用大语言模型（LLM）公平性的内在不可行性 large language model
20 LLMs and Memorization: On Quality and Specificity of Copyright Compliance 提出系统性分析方法，评估大型语言模型在版权合规性方面的表现，并分析其规避行为。 large language model ✅
21 The Battle of LLMs: A Comparative Study in Conversational QA Tasks 对比研究：大型语言模型在对话式问答任务中的性能评估 large language model
22 MockLLM: A Multi-Agent Behavior Collaboration Framework for Online Job Seeking and Recruiting 提出MockLLM，用于在线招聘中模拟面试交互，提升人岗匹配精度。 large language model
23 Spanish and LLM Benchmarks: is MMLU Lost in Translation? 揭示MMLU基准测试翻译陷阱：西班牙语场景下LLM性能评估的挑战与改进 large language model
24 fMRI predictors based on language models of increasing complexity recover brain left lateralization 利用不同复杂度的语言模型预测fMRI，揭示大脑左侧优势 large language model
25 Recent Trends in Personalized Dialogue Generation: A Review of Datasets, Methodologies, and Evaluations 综述个性化对话生成：数据集、方法与评估的最新趋势 large language model
26 XL3M: A Training-free Framework for LLM Length Extension Based on Segment-wise Inference XL3M：一种基于分段推理的LLM长度泛化零训练框架 large language model
27 Recent Advances of Foundation Language Models-based Continual Learning: A Survey 综述：基于大语言模型的持续学习研究进展与方法分类 large language model
28 Decoding moral judgement from text: a pilot study 探索基于文本的道德判断解码：一项脑机接口的初步研究 large language model
29 Thai Winograd Schemas: A Benchmark for Thai Commonsense Reasoning 提出泰语Winograd Schema基准测试，用于评估泰语常识推理能力 large language model
30 Semantic are Beacons: A Semantic Perspective for Unveiling Parameter-Efficient Fine-Tuning in Knowledge Learning 针对知识学习，提出基于语义感知的参数高效微调方法 large language model
31 Aligning to Thousands of Preferences via System Message Generalization 提出Janus，通过系统消息泛化实现LLM对用户个性化偏好的对齐 large language model ✅
32 More Than Catastrophic Forgetting: Integrating General Capabilities For Domain-Specific LLMs 提出ALoRA，解决领域LLM通用能力集成难题，提升领域任务性能 large language model
33 Detection-Correction Structure via General Language Model for Grammatical Error Correction 提出基于通用语言模型GLM的检测-纠错结构DeCoGLM，用于语法纠错。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
34 Learning diverse attacks on large language models for robust red-teaming and safety tuning 提出基于GFlowNet的大语言模型多样性攻击方法，提升红队测试和安全调优效果 reinforcement learning large language model
35 SLMRec: Distilling Large Language Models into Small for Sequential Recommendation 提出SLMRec，通过知识蒸馏将大语言模型压缩为小模型用于序列推荐 distillation large language model
36 Enhancing Emotion Recognition in Conversation through Emotional Cross-Modal Fusion and Inter-class Contrastive Learning 提出基于跨模态融合和类间对比学习的对话情感识别增强方法 contrastive learning motion prediction
37 Online Merging Optimizers for Boosting Rewards and Mitigating Tax in Alignment 提出在线融合优化器，提升奖励并缓解对齐中的能力退化问题 reinforcement learning RLHF DPO
38 Are PPO-ed Language Models Hackable? 研究PPO对齐的语言模型是否易受攻击，并探索对抗方法 PPO
39 ORLM: A Customizable Framework in Training Large Models for Automated Optimization Modeling 提出ORLM框架以解决优化建模自动化问题 reinforcement learning large language model ✅
40 Modeling Dynamic Topics in Chain-Free Fashion by Evolution-Tracking Contrastive Learning and Unassociated Word Exclusion 提出链式无关动态主题模型，通过进化追踪对比学习和非相关词排除解决主题漂移问题。 contrastive learning ✅
41 LeDex: Training LLMs to Better Self-Debug and Explain Code LeDex：训练LLM以提升代码自调试和解释能力 reinforcement learning reward design

🔬 支柱七：动作重定向 (Motion Retargeting) (1 篇)

#题目一句话要点标签🔗⭐
42 Long Context is Not Long at All: A Prospector of Long-Dependency Data for Large Language Models 提出ProLong框架，挖掘长依赖数据以提升大语言模型长文本建模能力。 spatial relationship large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
43 Interpretable classification of wiki-review streams 提出一种可解释的Wiki评论流分类方法，用于预测和解释应回退的评论。 manipulation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页