cs.CL（2025-05-08）

📊 共 24 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (19 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (5)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (19 篇)

#题目一句话要点标签🔗⭐
1 Do MLLMs Capture How Interfaces Guide User Behavior? A Benchmark for Multimodal UI/UX Design Understanding 提出WiserUI-Bench基准，评估MLLM在理解UI/UX设计对用户行为影响方面的能力 large language model multimodal
2 Chain-of-Thought Tokens are Computer Program Variables 研究表明CoT中的Token类似于程序变量，可有效解决复杂推理任务 large language model chain-of-thought ✅
3 Toward Reasonable Parrots: Why Large Language Models Should Argue with Us by Design 设计具备论证能力的“合理鹦鹉”型大语言模型，提升批判性思维能力 large language model
4 A Benchmark Dataset and a Framework for Urdu Multimodal Named Entity Recognition 提出U-MNER框架与Twitter2015-Urdu数据集，推进乌尔都语多模态命名实体识别研究。 multimodal
5 Unveiling Language-Specific Features in Large Language Models via Sparse Autoencoders 利用稀疏自编码器揭示大型语言模型中的语言特定特征 large language model ✅
6 Performance Evaluation of Large Language Models in Bangla Consumer Health Query Summarization 评估大型语言模型在孟加拉语消费者健康查询摘要任务中的性能 large language model
7 Scalable Multi-Stage Influence Function for Large Language Models via Eigenvalue-Corrected Kronecker-Factored Parameterization 提出基于EK-FAC的大规模多阶段影响函数，用于分析微调LLM对预训练数据的依赖 large language model ✅
8 Bring Reason to Vision: Understanding Perception and Reasoning through Model Merging 通过模型融合，将大型语言模型的推理能力迁移至视觉-语言模型 large language model multimodal
9 Crosslingual Reasoning through Test-Time Scaling 通过测试时缩放提升英语中心语言模型跨语言推理能力 large language model chain-of-thought
10 KG-HTC: Integrating Knowledge Graphs into LLMs for Effective Zero-shot Hierarchical Text Classification 提出KG-HTC，通过融合知识图谱与LLM，有效解决零样本分层文本分类问题。 large language model ✅
11 ComPO: Preference Alignment via Comparison Oracles 提出ComPO，通过比较Oracle进行偏好对齐，解决LLM中的噪声偏好问题 large language model
12 UKElectionNarratives: A Dataset of Misleading Narratives Surrounding Recent UK General Elections 构建英国大选误导性叙事数据集，并评估大型语言模型检测能力。 large language model
13 clem:todd: A Framework for the Systematic Benchmarking of LLM-Based Task-Oriented Dialogue System Realisations clem todd：用于系统评测基于LLM的任务型对话系统实现的框架 large language model
14 Ultra-FineWeb: Efficient Data Filtering and Verification for High-Quality LLM Training Data Ultra-FineWeb：高效数据过滤与验证，提升大语言模型训练数据质量 large language model
15 Frame In, Frame Out: Do LLMs Generate More Biased News Headlines than Humans? 研究表明，大型语言模型比人类更易生成带有偏见的新闻标题。 large language model
16 RICo: Refined In-Context Contribution for Automatic Instruction-Tuning Data Selection 提出RICo，通过上下文学习改进指令微调数据选择，提升大模型性能。 large language model
17 Product of Experts with LLMs: Boosting Performance on ARC Is a Matter of Perspective 利用专家乘积与LLM提升ARC性能：视角是关键 large language model
18 Reliably Bounding False Positives: A Zero-Shot Machine-Generated Text Detection Framework via Multiscaled Conformal Prediction 提出基于多尺度共形预测的零样本机器生成文本检测框架，可靠控制误报率。 large language model
19 Rethinking Invariance in In-context Learning 提出InvICL，解决上下文学习中对示例顺序敏感且现有不变方法性能不足的问题。 large language model ✅

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
20 Enhanced Urdu Intent Detection with Large Language Models and Prototype-Informed Predictive Pipelines 提出LLMPIA框架，提升大型语言模型在乌尔都语意图检测中的性能 representation learning contrastive learning large language model
21 Latent Preference Coding: Aligning Large Language Models via Discrete Latent Codes 提出隐式偏好编码(LPC)，通过离散隐变量对齐大语言模型，提升对人类偏好的建模能力。 DPO large language model
22 Scalable LLM Math Reasoning Acceleration with Low-rank Distillation Caprese：低秩蒸馏加速LLM数学推理，显著降低计算成本 distillation large language model
23 Scaling Laws for Speculative Decoding 针对思辨解码，提出Log-linear Scaling Laws，加速LLM推理。 RLHF large language model chain-of-thought
24 Reasoning Models Don't Always Say What They Think 评估思维链模型的忠实性，揭示其推理过程与实际行为的不一致性 reinforcement learning chain-of-thought

⬅️ 返回 cs.CL 首页 · 🏠 返回主页