cs.CL（2025-04-07）

📊 共 30 篇论文 | 🔗 7 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (26 🔗5) 支柱二：RL算法与架构 (RL & Architecture) (4 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (26 篇)

#题目一句话要点标签🔗⭐
1 SEAL: Steerable Reasoning Calibration of Large Language Models for Free SEAL：一种免训练的可操纵推理校准方法，提升大语言模型推理效率与准确率 large language model chain-of-thought ✅
2 Can Large Language Models Match Tutoring System Adaptivity? A Benchmarking Study 评估大语言模型在智能辅导系统中的适应性：一项基准研究 large language model instruction following
3 A Survey on Hypothesis Generation for Scientific Discovery in the Era of Large Language Models 综述：大型语言模型时代科学发现的假设生成方法 large language model multimodal
4 The Curse of CoT: On the Limitations of Chain-of-Thought in In-Context Learning CoT在基于模式的上下文学习中表现欠佳：揭示其局限性 large language model chain-of-thought
5 Evaluating the Generalization Capabilities of Large Language Models on Code Reasoning 评估大型语言模型在代码推理上的泛化能力 large language model
6 Less but Better: Parameter-Efficient Fine-Tuning of Large Language Models for Personality Detection PersLLM：面向人格检测，高效微调大语言模型的参数 large language model
7 Leveraging Large Language Models for Cost-Effective, Multilingual Depression Detection and Severity Assessment 利用大型语言模型进行高性价比、多语种的抑郁症检测与严重程度评估 large language model
8 'Neural howlround' in large language models: a self-reinforcing bias phenomenon, and a dynamic attenuation solution 提出神经啸叫衰减机制，解决大语言模型中自增强偏差导致的推理失效问题 large language model
9 Do Large Language Models Truly Grasp Addition? A Rule-Focused Diagnostic Using Two-Integer Arithmetic 通过规则诊断，揭示大语言模型在二整数加法中依赖模式匹配而非真正理解 large language model ✅
10 A Domain-Based Taxonomy of Jailbreak Vulnerabilities in Large Language Models 提出基于领域的大语言模型越狱漏洞分类法，提升对模型安全风险的理解。 large language model
11 Quantization Hurts Reasoning? An Empirical Study on Quantized Reasoning Models 系统研究量化对推理语言模型的影响，揭示模型大小、来源和任务难度是关键因素。 large language model chain-of-thought ✅
12 COIG-P: A High-Quality and Large-Scale Chinese Preference Dataset for Alignment with Human Values COIG-P：一个高质量、大规模的中文偏好数据集，用于对齐人类价值观 large language model multimodal ✅
13 Truthful or Fabricated? Using Causal Attribution to Mitigate Reward Hacking in Explanations 利用因果归因缓解解释中的奖励欺骗问题 large language model chain-of-thought
14 TathyaNyaya and FactLegalLlama: Advancing Factual Judgment Prediction and Explanation in the Indian Legal Context 提出TathyaNyaya数据集和FactLegalLlama模型，用于印度法律背景下的事实性判决预测与解释。 large language model
15 Are You Getting What You Pay For? Auditing Model Substitution in LLM APIs 提出基于TEE的LLM API模型替换检测方案，保障用户权益。 large language model ✅
16 Bridging Industrial Expertise and XR with LLM-Powered Conversational Agents 提出基于LLM的XR工业知识助手，解决工业领域知识传递难题。 large language model
17 Following the Whispers of Values: Unraveling Neural Mechanisms Behind Value-Oriented Behaviors in LLMs 提出ValueExploration框架，探索LLM中价值观驱动行为的神经机制 large language model
18 Pretraining Language Models for Diachronic Linguistic Change Discovery 提出一种高效的领域限制预训练方法，用于发现历时语言变化。 large language model
19 Enhancing LLM-Based Short Answer Grading with Retrieval-Augmented Generation 提出基于检索增强生成(RAG)的LLM短答案自动评分框架，提升评分准确性 large language model
20 LLM-based Automated Grading with Human-in-the-Loop 提出GradeHITL框架，利用人机协同提升LLM自动评分的准确性 large language model
21 Enhancing NER Performance in Low-Resource Pakistani Languages using Cross-Lingual Data Augmentation 提出跨语言数据增强方法，提升低资源巴基斯坦语种命名实体识别性能 large language model
22 Not All Data Are Unlearned Equally 揭示LLM中数据频率对不可学习性的影响，强调评估与方法改进 large language model
23 Voices of Freelance Professional Writers on AI: Limitations, Expectations, and Fears 调查AI对自由撰稿人的影响：局限性、期望与担忧 large language model
24 Improving Multilingual Retrieval-Augmented Language Models through Dialectic Reasoning Argumentations 提出Dialectic-RAG，通过辩证推理增强多语言检索增强语言模型，提升知识利用和鲁棒性。 large language model
25 Can LLMs Interpret and Leverage Structured Linguistic Representations? A Case Study with AMRs 利用AMR结构化信息，提升LLM在长文本任务中的性能 large language model
26 Sequential-NIAH: A Needle-In-A-Haystack Benchmark for Extracting Sequential Needles from Long Contexts 提出Sequential-NIAH基准，评估LLM从长文本中提取序列信息的能力 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🔗⭐
27 Beyond Single-Turn: A Survey on Multi-Turn Interactions with Large Language Models 综述LLM多轮交互：评估、增强方法及未来方向，覆盖多领域任务。 reinforcement learning large language model instruction following ✅
28 Revealing the Intrinsic Ethical Vulnerability of Aligned Large Language Models 揭示对齐大语言模型内在的伦理脆弱性 preference learning large language model
29 Concise Reasoning via Reinforcement Learning 提出基于强化学习的两阶段训练方法，显著减少推理模型token使用量并保持或提升准确率。 reinforcement learning PPO
30 CARE: Multilingual Human Preference Learning for Cultural Awareness CARE：通过多语言人类偏好学习提升语言模型的文化感知能力 preference learning ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页