cs.CL（2024-06-06）

📊 共 48 篇论文 | 🔗 9 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (41 🔗7) 支柱二：RL算法与架构 (RL & Architecture) (6 🔗1) 支柱五：交互与反应 (Interaction & Reaction) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (41 篇)

#题目一句话要点标签🔗⭐
1 XL-HeadTags: Leveraging Multimodal Retrieval Augmentation for the Multilingual Generation of News Headlines and Tags 提出XL-HeadTags，利用多模态检索增强实现新闻标题和标签的多语言生成。 multimodal
2 PaCE: Parsimonious Concept Engineering for Large Language Models PaCE：用于大语言模型的简约概念工程，通过激活空间操作实现模型对齐。 large language model
3 Pointer-Guided Pre-Training: Infusing Large Language Models with Paragraph-Level Contextual Awareness 提出指针引导预训练方法，增强大语言模型对段落级上下文的理解能力 large language model
4 Legal Documents Drafting with Fine-Tuned Pre-Trained Large Language Model 利用微调预训练大语言模型进行法律文书起草 large language model
5 Confabulation: The Surprising Value of Large Language Model Hallucinations 重新审视大语言模型幻觉：将其视为提升叙事性和连贯性的潜在资源 large language model
6 Ask LLMs Directly, "What shapes your bias?": Measuring Social Bias in Large Language Models 提出直接询问LLM偏见来源的方法，量化评估大语言模型中的社会偏见 large language model
7 Are Large Language Models the New Interface for Data Pipelines? 探索大型语言模型作为数据管道新界面的潜力 large language model
8 llmNER: (Zero|Few)-Shot Named Entity Recognition, Exploiting the Power of Large Language Models llmNER：利用大型语言模型实现零样本和少样本命名实体识别 large language model
9 Buffer of Thoughts: Thought-Augmented Reasoning with Large Language Models 提出Buffer of Thoughts (BoT)，增强LLM推理的准确性、效率和鲁棒性。 large language model ✅
10 Benchmark Data Contamination of Large Language Models: A Survey 综述大型语言模型基准数据污染问题，并探讨缓解策略与未来方向 large language model
11 ValueBench: Towards Comprehensively Evaluating Value Orientations and Understanding of Large Language Models 提出ValueBench以全面评估大语言模型的价值取向 large language model ✅
12 Uncovering Limitations of Large Language Models in Information Seeking from Tables 提出TabIS基准，揭示大语言模型在表格信息检索中的局限性 large language model
13 Spontaneous Speech-Based Suicide Risk Detection Using Whisper and Large Language Models 利用Whisper和LLM进行基于语音的青少年自杀风险预测 large language model
14 Performance of large language models in numerical vs. semantic medical knowledge: Benchmarking on evidence-based Q&As EBMQA基准测试揭示大语言模型在数值医学知识处理上弱于语义知识，且逊于人类专家 large language model
15 Chaos with Keywords: Exposing Large Language Models Sycophantic Hallucination to Misleading Keywords and Evaluating Defense Strategies 揭示大语言模型在误导性关键词下的谄媚性幻觉并评估防御策略 large language model
16 A Survey on Medical Large Language Models: Technology, Application, Trustworthiness, and Future Directions 医学大语言模型综述：技术、应用、可信赖性与未来方向 large language model
17 Improving Audio Codec-based Zero-Shot Text-to-Speech Synthesis with Multi-Modal Context and Large Language Model 提出MMCE-Qformer，利用多模态上下文和LLM改进基于音频编解码器的零样本TTS，适用于长文本语音合成。 large language model
18 MAIRA-2: Grounded Radiology Report Generation MAIRA-2：提出基于定位信息的放射报告生成模型，提升报告质量与可验证性。 large language model multimodal
19 A + B: A General Generator-Reader Framework for Optimizing LLMs to Unleash Synergy Potential 提出A+B通用生成-阅读框架，优化LLM以释放协同潜力 large language model foundation model
20 BLSP-Emo: Towards Empathetic Large Speech-Language Models 提出BLSP-Emo，一种支持情感理解的端到端语音-语言预训练模型 multimodal instruction following
21 Light-PEFT: Lightening Parameter-Efficient Fine-Tuning via Early Pruning Light-PEFT：通过早期剪枝减轻参数高效微调的计算负担 large language model foundation model
22 LLMEmbed: Rethinking Lightweight LLM's Genuine Function in Text Classification 提出LLMEmbed，一种轻量级LLM文本分类迁移学习策略，提升效率与性能。 large language model chain-of-thought ✅
23 ArMeme: Propagandistic Content in Arabic Memes 构建阿拉伯语宣传模因数据集，为检测多模态有害信息提供资源。 multimodal
24 Time Sensitive Knowledge Editing through Efficient Finetuning 提出基于高效微调的时间敏感知识编辑方法，提升LLM知识更新能力。 large language model
25 Reinterpreting 'the Company a Word Keeps': Towards Explainable and Ontologically Grounded Language Models 提出一种可解释且本体论证的语言模型，旨在克服现有大语言模型的局限性。 large language model
26 Transformers need glasses! Information over-squashing in language tasks 揭示Transformer语言模型的信息过挤压问题，并提出潜在解决方案 large language model
27 What Do Language Models Learn in Context? The Structured Task Hypothesis 通过组合预训练任务，语言模型可实现上下文学习 large language model
28 Do Language Models Understand Morality? Towards a Robust Detection of Moral Content 提出基于大语言模型的零样本道德内容检测方法，提升跨领域鲁棒性 large language model
29 Legal Judgment Reimagined: PredEx and the Rise of Intelligent AI Interpretation in Indian Courts 提出PredEx数据集，提升LLM在印度法律判决预测与解释中的准确性 large language model
30 Automatic Bug Detection in LLM-Powered Text-Based Games Using LLMs 提出基于LLM的自动方法，用于检测LLM驱动的文本游戏中存在的缺陷。 large language model
31 MoralBench: Moral Evaluation of LLMs MoralBench：用于评估大型语言模型道德推理能力的新基准 large language model ✅
32 BEADs: Bias Evaluation Across Domains 提出BEADs数据集，用于跨领域偏见评估，促进负责任的AI系统开发。 large language model ✅
33 Every Answer Matters: Evaluating Commonsense with Probabilistic Measures 提出常识框架补全任务CFC与概率评估方法，解决常识理解中多答案与偏见问题。 large language model
34 Exploring the Latest LLMs for Leaderboard Extraction 探索大型语言模型在AI研究论文排行榜信息抽取中的应用 large language model
35 Evaluating the Smooth Control of Attribute Intensity in Text Generation with LLMs 提出平滑控制文本生成属性强度的方法以解决生成一致性问题 large language model ✅
36 What Languages are Easy to Language-Model? A Perspective from Learning Probabilistic Regular Languages 研究RNN和Transformer语言模型学习概率正则语言的能力，揭示影响学习的关键因素。 large language model
37 Characterizing Similarities and Divergences in Conversational Tones in Humans and LLMs by Sampling with People 提出一种迭代采样方法，用于表征人类与LLM在对话语气上的异同 large language model
38 DICE: Detecting In-distribution Contamination in LLM's Fine-tuning Phase for Math Reasoning 提出DICE以检测大语言模型微调阶段的数据污染问题 large language model ✅
39 Assessing LLMs for Zero-shot Abstractive Summarization Through the Lens of Relevance Paraphrasing 提出相关性释义方法，评估大型语言模型在零样本摘要生成中的鲁棒性 large language model
40 Effective Context Selection in LLM-based Leaderboard Generation: An Empirical Study 提出基于LLM的上下文选择方法，高效生成AI研究排行榜，提升准确率并减少幻觉。 large language model
41 NAP^2: A Benchmark for Naturalness and Privacy-Preserving Text Rewriting by Learning from Human NAP^2：通过模仿人类行为，实现自然且保护隐私的文本重写基准。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
42 M-QALM: A Benchmark to Assess Clinical Reading Comprehension and Knowledge Recall in Large Language Models via Question Answering M-QALM：构建临床阅读理解与知识回忆问答基准，评估大型语言模型能力 representation learning large language model
43 Prototypical Reward Network for Data-Efficient RLHF 提出原型奖励网络以解决数据效率低下的问题 reinforcement learning RLHF large language model
44 Evaluating the World Model Implicit in a Generative Model 提出基于Myhill-Nerode定理的评估指标，揭示生成模型中隐式世界模型的脆弱性 world model large language model
45 UltraMedical: Building Specialized Generalists in Biomedicine UltraMedical：构建生物医学领域的专业通用大语言模型 reinforcement learning preference learning direct preference optimization ✅
46 Proofread: Fixes All Errors with One Tap Proofread：利用大语言模型，一键修复Gboard中的句子和段落错误 reinforcement learning large language model
47 Speculative Decoding via Early-exiting for Faster LLM Inference with Thompson Sampling Control Mechanism 提出基于早退机制和Thompson采样的推测解码方法，加速LLM推理。 distillation large language model

🔬 支柱五：交互与反应 (Interaction & Reaction) (1 篇)

#题目一句话要点标签🔗⭐
48 Lean Workbook: A large-scale Lean problem set formalized from natural language math problems 提出Lean Workbook，用于解决形式化语言数学定理证明中训练数据稀缺问题 IMoS large language model ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页