cs.CL（2025-02-07）

📊 共 30 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (25 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗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) (25 篇)

#题目一句话要点标签🔗⭐
1 Survey on Vision-Language-Action Models AI辅助文献综述：探索大型语言模型在视觉-语言-动作模型研究中的应用 vision-language-action VLA large language model
2 Transforming Science with Large Language Models: A Survey on AI-assisted Scientific Discovery, Experimentation, Content Generation, and Evaluation 综述：大型语言模型赋能科学研究，涵盖AI辅助发现、实验、内容生成与评估 large language model multimodal
3 Beyond External Monitors: Enhancing Transparency of Large Language Models for Easier Monitoring 提出TELLME方法，提升大语言模型透明度，便于监控不当行为 large language model chain-of-thought
4 M-IFEval: Multilingual Instruction-Following Evaluation 提出M-IFEval多语言指令跟随评估基准，扩展LLM评估至法语、日语和西班牙语。 large language model instruction following
5 Concept Navigation and Classification via Open-Source Large Language Model Processing 提出一种基于开源大语言模型的概念导航与分类框架，用于文本数据中潜在结构的检测与分类。 large language model
6 ARR: Question Answering with Large Language Models via Analyzing, Retrieving, and Reasoning 提出ARR框架，通过分析、检索和推理增强大语言模型在问答任务中的性能 large language model
7 Evaluating Personality Traits in Large Language Models: Insights from Psychological Questionnaires 利用心理问卷评估大型语言模型的人格特质，揭示其内在性格差异 large language model
8 Probabilistic Subspace Manifolds for Contextual Inference in Large Language Models 提出基于概率子空间流形的LLM上下文推断方法，提升语义粒度和鲁棒性 large language model
9 Probing Internal Representations of Multi-Word Verbs in Large Language Models 探究大型语言模型中多词动词内部表征，揭示词汇和句法属性的编码方式。 large language model
10 S$^2$-MAD: Breaking the Token Barrier to Enhance Multi-Agent Debate Efficiency S$^2$-MAD：突破Token限制，提升多智能体辩论效率 large language model chain-of-thought
11 SeDi-Instruct: Enhancing Alignment of Language Models through Self-Directed Instruction Generation 提出SeDi-Instruct，通过自定向指令生成提升语言模型对齐效果。 large language model foundation model
12 Self-Supervised Prompt Optimization 提出自监督提示优化（SPO）框架，无需外部参考即可提升LLM在各类任务中的性能。 large language model ✅
13 Aligning Black-box Language Models with Human Judgments 提出线性映射框架，对齐黑盒语言模型与人类主观判断 large language model
14 LLM-Supported Natural Language to Bash Translation 提出高质量数据集与评估方法，提升LLM在自然语言到Bash命令翻译中的准确率。 large language model ✅
15 Flexible and Efficient Grammar-Constrained Decoding 提出一种更快语法约束解码算法，加速LLM生成结构化输出。 large language model
16 GSM-Infinite: How Do Your LLMs Behave over Infinitely Increasing Context Length and Reasoning Complexity? 提出GSM-Infinite，用于评估LLM在无限增长的上下文长度和推理复杂性下的表现 large language model
17 Enhancing Knowledge Graph Construction: Evaluating with Emphasis on Hallucination, Omission, and Graph Similarity Metrics 提出基于BERTScore的知识图谱构建评估框架，关注幻觉和遗漏问题 large language model
18 Extracting and Understanding the Superficial Knowledge in Alignment 提出一种提取和理解对齐模型中浅层知识的方法，用于高效模型对齐和安全恢复。 large language model
19 Fine-Tuned LLMs are "Time Capsules" for Tracking Societal Bias Through Books 利用微调LLM和书籍追踪社会偏见随时间演变 large language model
20 Can LLMs Rank the Harmfulness of Smaller LLMs? We are Not There Yet 评估大型语言模型标注小型语言模型有害内容的能力，结果表明现有方法尚不成熟。 large language model
21 NoLiMa: Long-Context Evaluation Beyond Literal Matching NoLiMa：提出超越字面匹配的长文本评估基准，揭示LLM在长程推理中的局限性 large language model ✅
22 CodeSCM: Causal Analysis for Multi-Modal Code Generation 提出CodeSCM，用于多模态代码生成中因果效应分析。 large language model
23 An Annotated Reading of 'The Singer of Tales' in the LLM Era 利用大型语言模型视角解读《故事的歌手》，探索口头诗歌创作理论 large language model
24 nvAgent: Automated Data Visualization from Natural Language via Collaborative Agent Workflow 提出nvAgent，通过协同Agent工作流解决复杂NL2Vis任务中多表推理难题。 large language model
25 Developmentally-plausible Working Memory Shapes a Critical Period for Language Acquisition 提出一种发展可信的工作记忆模型，加速语言模型在关键期的学习效率 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
26 LLMs Can Teach Themselves to Better Predict the Future 提出基于自博弈和直接偏好优化的框架，提升LLM的未来预测能力 DPO direct preference optimization large language model
27 SSMLoRA: Enhancing Low-Rank Adaptation with State Space Model 提出SSMLoRA，利用状态空间模型增强低秩适应，提升参数效率。 SSM state space model ✅
28 Before It's Too Late: A State Space Model for the Early Prediction of Misinformation and Disinformation Engagement 提出IC-Mamba模型，用于社交媒体上虚假信息传播的早期预测。 Mamba state space model ✅
29 DuoGuard: A Two-Player RL-Driven Framework for Multilingual LLM Guardrails 提出DuoGuard框架以解决多语言LLM安全性问题 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🔗⭐
30 Commonality and Individuality! Integrating Humor Commonality with Speaker Individuality for Humor Recognition 提出CIHR模型，融合幽默共性和说话者个性以提升幽默识别效果 HuMoR

⬅️ 返回 cs.CL 首页 · 🏠 返回主页