cs.CL（2025-09-02）

📊 共 34 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) (24 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗1) 支柱七：动作重定向 (Motion Retargeting) (1 🔗1) 支柱四：生成式动作 (Generative Motion) (1) 支柱三：空间感知与语义 (Perception & Semantics) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 VaccineRAG: Boosting Multimodal Large Language Models' Immunity to Harmful RAG Samples VaccineRAG：通过对抗有害RAG样本提升多模态大语言模型的免疫力 large language model multimodal chain-of-thought
2 IDEAlign: Comparing Large Language Models to Human Experts in Open-ended Interpretive Annotations IDEAlign：通过“奇数挑一”评估LLM在开放式解释性标注任务中与人类专家的对齐程度 large language model
3 Clustering Discourses: Racial Biases in Short Stories about Women Generated by Large Language Models 利用LLaMA 3.2-3B生成短篇小说，揭示其中关于黑人女性和白人女性的种族偏见。 large language model
4 Scaling behavior of large language models in emotional safety classification across sizes and tasks 研究LLM在情感安全分类中的规模效应，探索轻量级模型在心理健康领域的应用潜力。 large language model
5 Comparative Study of Pre-Trained BERT and Large Language Models for Code-Mixed Named Entity Recognition 对比研究预训练BERT与大语言模型在混合语命名实体识别中的应用 large language model
6 An Ensemble Classification Approach in A Multi-Layered Large Language Model Framework for Disease Prediction 提出一种多层LLM框架下的集成方法，用于提升阿拉伯语社交健康数据中的疾病预测精度。 large language model
7 E-THER: A Multimodal Dataset for Empathic AI -- Towards Emotional Mismatch Awareness 提出E-THER多模态数据集，用于提升AI在情感不匹配感知方面的能力。 multimodal
8 DeepSeek performs better than other Large Language Models in Dental Cases DeepSeek在大语言模型牙科病例分析中表现优于其他模型 large language model
9 Behavioral Fingerprinting of Large Language Models 提出大语言模型行为指纹框架，用于剖析模型认知与交互风格的差异。 large language model ✅
10 DRAssist: Dispute Resolution Assistance using Large Language Models DRAssist：利用大型语言模型辅助解决汽车保险和域名争议 large language model
11 Extracting OPQRST in Electronic Health Records using Large Language Models with Reasoning 提出基于LLM的推理方法，用于从电子病历中提取OPQRST信息，提升临床决策效率。 large language model
12 FActBench: A Benchmark for Fine-grained Automatic Evaluation of LLM-Generated Text in the Medical Domain 构建医学领域LLM生成文本自动评估基准FActBench，提升事实性评估的可靠性 large language model chain-of-thought
13 How Instruction-Tuning Imparts Length Control: A Cross-Lingual Mechanistic Analysis 指令调优通过专门化模型深层组件实现跨语言长度控制 large language model foundation model
14 PalmX 2025: The First Shared Task on Benchmarking LLMs on Arabic and Islamic Culture PalmX 2025：首个面向阿拉伯和伊斯兰文化的大语言模型评测共享任务 large language model
15 MoSEs: Uncertainty-Aware AI-Generated Text Detection via Mixture of Stylistics Experts with Conditional Thresholds MoSEs：通过混合文体专家模型与条件阈值实现不确定性感知的AI生成文本检测 large language model ✅
16 SpecEval: Evaluating Model Adherence to Behavior Specifications SpecEval：评估大型模型对行为规范的遵循程度 foundation model
17 LLMs and their Limited Theory of Mind: Evaluating Mental State Annotations in Situated Dialogue 提出双步框架，利用LLM评估团队对话中共享心智模型的偏差。 large language model
18 Towards Fundamental Language Models: Does Linguistic Competence Scale with Model Size? 提出基础语言模型范式，探索语言能力与模型规模的解耦策略 large language model
19 Avoidance Decoding for Diverse Multi-Branch Story Generation 提出Avoidance Decoding，解决LLM故事生成中多样性不足和重复性问题。 large language model
20 AMBEDKAR-A Multi-level Bias Elimination through a Decoding Approach with Knowledge Augmentation for Robust Constitutional Alignment of Language Models 提出AMBEDKAR框架，通过知识增强解码消除LLM中的多层次偏见，实现对印度宪法的稳健对齐。 large language model
21 JudgeAgent: Beyond Static Benchmarks for Knowledge-Driven and Dynamic LLM Evaluation 提出JudgeAgent，用于知识驱动和动态的大语言模型评估，突破静态基准限制 large language model ✅
22 Better by Comparison: Retrieval-Augmented Contrastive Reasoning for Automatic Prompt Optimization 提出CRPO：一种检索增强的对比推理框架，用于自动提示优化。 large language model
23 Attributes as Textual Genes: Leveraging LLMs as Genetic Algorithm Simulators for Conditional Synthetic Data Generation 提出Genetic Prompt，利用LLM模拟遗传算法进行条件合成数据生成，提升数据质量和多样性。 large language model
24 Context Engineering for Trustworthiness: Rescorla Wagner Steering Under Mixed and Inappropriate Contexts 提出RW-Steering，解决LLM在混合不当上下文中易受少量有害信息影响的问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (7 篇)

#题目一句话要点标签🔗⭐
25 Understanding Reinforcement Learning for Model Training, and future directions with GRAPE 针对LLM指令调优，系统性解析强化学习算法并提出GRAPE新方向 reinforcement learning PPO DPO
26 Jointly Reinforcing Diversity and Quality in Language Model Generations 提出DARLING框架，联合强化语言模型生成的多样性和质量，提升创造性任务表现。 reinforcement learning large language model instruction following
27 Implicit Actor Critic Coupling via a Supervised Learning Framework for RLVR 提出PACS框架，通过监督学习隐式耦合Actor-Critic，提升RLVR中LLM的推理能力。 reinforcement learning PPO large language model ✅
28 ProST: Progressive Sub-task Training for Pareto-Optimal Multi-agent Systems Using Small Language Models 提出ProST渐进式子任务训练，提升小型语言模型多智能体系统在复杂任务中的效率和效果。 curriculum learning large language model
29 GRAM-R$^2$: Self-Training Generative Foundation Reward Models for Reward Reasoning 提出GRAM-R$^2$，通过自训练生成式奖励模型实现奖励推理，提升下游任务性能。 reinforcement learning foundation model
30 DCPO: Dynamic Clipping Policy Optimization DCPO：动态裁剪策略优化，提升LLM在可验证奖励下的推理能力 reinforcement learning large language model
31 StructCoh: Structured Contrastive Learning for Context-Aware Text Semantic Matching 提出StructCoh以解决文本语义匹配中的结构性和细粒度语义问题 contrastive learning

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

#题目一句话要点标签🔗⭐
32 Implicit Reasoning in Large Language Models: A Comprehensive Survey 综述：大型语言模型中的隐式推理研究进展与执行范式分析 latent optimization large language model chain-of-thought ✅

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
33 Spectrogram Patch Codec: A 2D Block-Quantized VQ-VAE and HiFi-GAN for Neural Speech Coding 提出一种基于2D块量化VQ-VAE和HiFi-GAN的神经语音编码方法，简化了残差量化流程。 VQ-VAE

🔬 支柱三：空间感知与语义 (Perception & Semantics) (1 篇)

#题目一句话要点标签🔗⭐
34 SSVD: Structured SVD for Parameter-Efficient Fine-Tuning and Benchmarking under Domain Shift in ASR 提出SSVD，通过结构化SVD微调提升ASR领域迁移性能与效率。 semantic mapping semantic map foundation model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页