cs.CL（2025-09-17）

📊 共 28 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) (23 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (4) 支柱七：动作重定向 (Motion Retargeting) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (23 篇)

#题目一句话要点标签🔗⭐
1 Early Stopping Chain-of-thoughts in Large Language Models 提出ES-CoT，通过提前停止CoT生成降低大语言模型推理成本 large language model chain-of-thought
2 Simulating a Bias Mitigation Scenario in Large Language Models 构建模拟框架，评估缓解大语言模型偏见的策略 large language model
3 Annotating Training Data for Conditional Semantic Textual Similarity Measurement using Large Language Models 利用大型语言模型重标注条件语义文本相似度训练数据 large language model ✅
4 Do Large Language Models Understand Word Senses? 评估大型语言模型对词义理解能力，并验证其在词义消歧任务上的有效性。 large language model
5 Large Language Models Discriminate Against Speakers of German Dialects 大型语言模型对德语方言使用者存在歧视性偏见 large language model
6 How Can Quantum Deep Learning Improve Large Language Models? 探索量子深度学习在提升大型语言模型适应性方面的潜力 large language model
7 AssoCiAm: A Benchmark for Evaluating Association Thinking while Circumventing Ambiguity AssoCiAm：提出一个用于评估联想思维并规避歧义的基准 large language model multimodal
8 Enhancing Time Awareness in Generative Recommendation 提出GRUT模型，通过时间感知提升生成式推荐效果 large language model TAMP ✅
9 Estimating Semantic Alphabet Size for LLM Uncertainty Quantification 提出改进的语义字母表大小估计器，提升LLM不确定性量化的准确性和可解释性 large language model
10 Ticket-Bench: A Kickoff for Multilingual and Regionalized Agent Evaluation Ticket-Bench：多语言区域化Agent评估基准，提升任务型Agent的文化适应性 large language model
11 Correct-Detect: Balancing Performance and Ambiguity Through the Lens of Coreference Resolution in LLMs 揭示LLM在共指消解中性能与歧义检测的权衡：Correct-Detect 框架 large language model
12 Causal-Counterfactual RAG: The Integration of Causal-Counterfactual Reasoning into RAG 提出因果-反事实RAG，将因果推理融入RAG以提升知识密集型任务性能。 large language model
13 Adding LLMs to the psycholinguistic norming toolbox: A practical guide to getting the most out of human ratings 提出一种利用大型语言模型增强心理语言学规范数据集的方法 large language model
14 Apertus: Democratizing Open and Compliant LLMs for Global Language Environments Apertus：构建开放、合规且支持全球语言环境的大语言模型 large language model
15 ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution ShinkaEvolve：提出一种高效、开源的程序演化框架，用于解决科学发现中的样本效率问题。 large language model
16 Enhancing Multi-Agent Debate System Performance via Confidence Expression 提出ConfMAD框架，通过置信度表达提升多智能体辩论系统性能 large language model
17 Hala Technical Report: Building Arabic-Centric Instruction & Translation Models at Scale 提出Hala模型以提升阿拉伯语指令与翻译任务的性能 instruction following
18 Do LLMs Align Human Values Regarding Social Biases? Judging and Explaining Social Biases with LLMs 评估大语言模型在社会偏见场景下的人类价值观对齐程度，并分析其解释能力 large language model
19 Characterizing Knowledge Graph Tasks in LLM Benchmarks Using Cognitive Complexity Frameworks 利用认知复杂性框架表征LLM基准测试中知识图谱任务 large language model
20 Exploring Data and Parameter Efficient Strategies for Arabic Dialect Identifications 探索数据与参数高效的阿拉伯语方言识别策略 large language model
21 Thinking in a Crowd: How Auxiliary Information Shapes LLM Reasoning 研究辅助信息对LLM推理的影响：有害信息会显著降低模型性能 large language model ✅
22 Implementing a Logical Inference System for Japanese Comparatives 提出ccg-jcomp：一个基于组合语义的日语比较句逻辑推理系统 large language model
23 DSPC: Dual-Stage Progressive Compression Framework for Efficient Long-Context Reasoning 提出DSPC双阶段渐进压缩框架，无需训练即可高效压缩长文本上下文。 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🔗⭐
24 Process-Supervised Reinforcement Learning for Interactive Multimodal Tool-Use Agents 提出Turn-level Adjudicated RL，解决交互式多模态工具使用Agent的训练难题 reinforcement learning large language model foundation model
25 Teaching According to Talents! Instruction Tuning LLMs with Competence-Aware Curriculum Learning 提出CAMPUS框架，通过能力感知课程学习提升指令调优大语言模型性能 curriculum learning large language model
26 SSL-SSAW: Self-Supervised Learning with Sigmoid Self-Attention Weighting for Question-Based Sign Language Translation 提出基于问题引导的自监督手语翻译框架SSL-SSAW，提升翻译质量。 contrastive learning multimodal
27 Noise Supervised Contrastive Learning and Feature-Perturbed for Anomalous Sound Detection 提出噪声监督对比学习和特征扰动方法，提升异常声音检测在同类型机器上的准确率。 contrastive learning

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

#题目一句话要点标签🔗⭐
28 Geometric Uncertainty for Detecting and Correcting Hallucinations in LLMs 提出基于几何不确定性的方法，用于检测和纠正LLM中的幻觉问题。 spatial relationship large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页