cs.CL（2025-09-12）

📊 共 25 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) (22 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1)

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

#题目一句话要点标签🔗⭐
1 A Survey on Retrieval And Structuring Augmented Generation with Large Language Models 综述检索与结构增强的大语言模型生成方法，解决幻觉、知识过时和领域受限问题。 large language model multimodal
2 Readme_AI: Dynamic Context Construction for Large Language Models Readme_AI：提出动态上下文构建方法，提升大语言模型在特定查询下的准确性和可靠性。 large language model ✅
3 Established Psychometric vs. Ecologically Valid Questionnaires: Rethinking Psychological Assessments in Large Language Models 对比心理测量问卷与生态效度问卷，重新评估大语言模型中的心理评估方法 large language model
4 Large Language Models Meet Legal Artificial Intelligence: A Survey 综述：大型语言模型赋能法律人工智能 large language model ✅
5 Arabic Large Language Models for Medical Text Generation 提出并微调阿拉伯语大型语言模型，用于生成医疗文本，辅助医院管理系统。 large language model
6 SearchInstruct: Enhancing Domain Adaptation via Retrieval-Based Instruction Dataset Creation SearchInstruct：通过检索增强的指令数据集创建提升领域自适应 large language model instruction following ✅
7 Pluralistic Alignment for Healthcare: A Role-Driven Framework 提出EthosAgents框架，增强医疗领域大语言模型对多元价值观的对齐。 large language model
8 Is In-Context Learning Learning? 研究表明上下文学习是一种有效的学习范式，但其泛化能力有限 chain-of-thought
9 Population-Aligned Persona Generation for LLM-based Social Simulation 提出人口对齐的Persona生成框架，提升LLM社会模拟的真实性和准确性 large language model
10 No Answer Needed: Predicting LLM Answer Accuracy from Question-Only Linear Probes 仅凭问题预测LLM答案准确性：线性探针揭示模型内部置信度 large language model
11 Benchmark of stylistic variation in LLM-generated texts 构建LLM文本风格基准：分析人类与AI生成文本的文体差异 large language model
12 On LLM-Based Scientific Inductive Reasoning Beyond Equations 提出SIRBench-V1基准，评估LLM在科学场景下超越方程的归纳推理能力 large language model
13 Opening the Black Box: Interpretable LLMs via Semantic Resonance Architecture 提出语义共振架构SRA，提升LLM可解释性与专家利用率。 large language model
14 Unsupervised Hallucination Detection by Inspecting Reasoning Processes 提出IRIS框架，通过检查LLM推理过程实现无监督幻觉检测 large language model
15 Reasoning Under Uncertainty: Exploring Probabilistic Reasoning Capabilities of LLMs 首个LLM概率推理能力综合研究：揭示优势与局限，探索未来改进方向 large language model
16 Context Copying Modulation: The Role of Entropy Neurons in Managing Parametric and Contextual Knowledge Conflicts 利用熵神经元抑制上下文复制，解决LLM参数知识与上下文冲突问题 large language model
17 RefactorCoderQA: Benchmarking LLMs for Multi-Domain Coding Question Solutions in Cloud and Edge Deployment 提出RefactorCoderQA基准和云边协同架构，提升LLM在多领域代码问题解决能力 large language model
18 Dropping Experts, Recombining Neurons: Retraining-Free Pruning for Sparse Mixture-of-Experts LLMs 提出DERN：一种免训练的专家剪枝与神经元重组框架，提升稀疏MoE LLM性能。 large language model
19 Beyond Token Limits: Assessing Language Model Performance on Long Text Classification 评估语言模型在长文本分类任务上的性能，发现长文本专用模型并无明显优势。 large language model
20 Incongruent Positivity: When Miscalibrated Positivity Undermines Online Supportive Conversations 研究表明，LLM在情感支持对话中易产生不恰当的积极回应，并提出检测方法。 large language model
21 HetaRAG: Hybrid Deep Retrieval-Augmented Generation across Heterogeneous Data Stores HetaRAG：跨异构数据存储的混合深度检索增强生成框架 large language model ✅
22 Whisper Has an Internal Word Aligner Whisper内部蕴含词对齐能力，无需额外训练即可高精度提取词级时间戳。 TAMP

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
23 From Correction to Mastery: Reinforced Distillation of Large Language Model Agents 提出SCoRe框架，通过强化蒸馏提升小模型Agent在复杂任务中的性能，媲美大模型。 reinforcement learning distillation large language model
24 SI-FACT: Mitigating Knowledge Conflict via Self-Improving Faithfulness-Aware Contrastive Tuning 提出SI-FACT框架，通过自提升的忠实度感知对比学习缓解LLM的知识冲突问题 contrastive learning large language model
25 DeepDive: Advancing Deep Search Agents with Knowledge Graphs and Multi-Turn RL DeepDive：利用知识图谱和多轮强化学习提升深度搜索Agent能力 reinforcement learning large language model ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页