cs.CL（2024-05-29）

📊 共 26 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (22 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱六：视频提取与匹配 (Video Extraction) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (22 篇)

#题目一句话要点标签🔗⭐
1 Weak-to-Strong Search: Align Large Language Models via Searching over Small Language Models 提出弱到强搜索方法，通过小模型搜索提升大语言模型的对齐效果。 large language model instruction following
2 Reverse Image Retrieval Cues Parametric Memory in Multimodal LLMs 利用反向图像检索提示增强多模态LLM的参数记忆，提升知识密集型任务性能 large language model multimodal
3 Unlocking the Potential of Large Language Models for Clinical Text Anonymization: A Comparative Study 提出基于大语言模型的临床文本匿名化方法以解决隐私保护问题 large language model
4 MAP-Neo: Highly Capable and Transparent Bilingual Large Language Model Series MAP-Neo：开源高性能透明双语大语言模型系列 large language model
5 Are Large Language Models Chameleons? An Attempt to Simulate Social Surveys 利用大型语言模型模拟社会调查，揭示文化、年龄和性别偏见 large language model
6 Unlearning Climate Misinformation in Large Language Models 研究气候虚假信息：评估并提升大语言模型的事实准确性 large language model
7 A Full-duplex Speech Dialogue Scheme Based On Large Language Models 提出基于LLM的全双工语音对话系统，显著降低对话延迟并提高交互精度。 large language model
8 PediatricsGPT: Large Language Models as Chinese Medical Assistants for Pediatric Applications 提出PediatricsGPT，构建中文儿科医疗大语言模型助手，提升诊断效率。 large language model
9 Evaluating the External and Parametric Knowledge Fusion of Large Language Models 提出知识融合评估框架，系统研究大语言模型外部知识与参数知识的融合能力 large language model
10 Towards Better Chain-of-Thought: A Reflection on Effectiveness and Faithfulness 深入分析CoT有效性和忠实性，提出信息增强算法提升推理性能 chain-of-thought
11 Prompting or Fine-tuning? Exploring Large Language Models for Causal Graph Validation 探索大语言模型在因果图验证中的应用：微调优于提示学习 large language model
12 Genshin: General Shield for Natural Language Processing with Large Language Models 提出Genshin框架，利用LLM作为防御插件提升NLP系统鲁棒性与可解释性 large language model
13 AlchemistCoder: Harmonizing and Eliciting Code Capability by Hindsight Tuning on Multi-source Data AlchemistCoder：通过多源数据上的后见之明调优，和谐并激发代码能力 large language model instruction following
14 CtrlA: Adaptive Retrieval-Augmented Generation via Inherent Control 提出CtrlA框架，通过内在控制实现自适应检索增强生成，提升LLM的诚实性和知识覆盖。 large language model ✅
15 Contextual Position Encoding: Learning to Count What's Important 提出上下文位置编码(CoPE)，解决LLM中基于上下文的位置寻址问题 large language model
16 Conveyor: Efficient Tool-aware LLM Serving with Tool Partial Execution Conveyor：通过工具部分执行提升LLM服务效率 large language model
17 Adaptive In-conversation Team Building for Language Model Agents 提出Captain Agent，自适应构建LLM Agent团队解决复杂任务 large language model
18 Nearest Neighbor Speculative Decoding for LLM Generation and Attribution 提出近邻推测解码（NEST），提升LLM生成质量、溯源能力和推理速度。 large language model ✅
19 Expert-Guided Extinction of Toxic Tokens for Debiased Generation 提出EXPOSED方法，通过专家引导消除LLM中的有害token，实现去偏见生成。 large language model
20 DGRC: An Effective Fine-tuning Framework for Distractor Generation in Chinese Multi-choice Reading Comprehension 提出DGRC框架，用于提升中文多项选择阅读理解中干扰项生成的质量。 chain-of-thought
21 MEMoE: Enhancing Model Editing with Mixture of Experts Adaptors 提出MEMoE，利用MoE适配器增强大语言模型的模型编辑能力 large language model
22 Language Generation with Strictly Proper Scoring Rules 提出基于严格Proper Scoring Rule的语言生成方法，提升模型生成能力 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🔗⭐
23 BLSP-KD: Bootstrapping Language-Speech Pre-training via Knowledge Distillation BLSP-KD：通过知识蒸馏引导语言-语音预训练，实现细粒度对齐 distillation large language model instruction following
24 Efficient Model-agnostic Alignment via Bayesian Persuasion 提出基于贝叶斯劝说的模型无关对齐框架，提升黑盒大模型性能 reinforcement learning RLHF large language model
25 Are You Sure? Rank Them Again: Repeated Ranking For Better Preference Datasets 提出重复排序法，提升AI反馈强化学习中偏好数据集质量 reinforcement learning large language model

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
26 Cracking the Code of Juxtaposition: Can AI Models Understand the Humorous Contradictions 提出YesBut基准，评估AI模型理解矛盾叙事漫画幽默的能力 HuMoR multimodal

⬅️ 返回 cs.CL 首页 · 🏠 返回主页