cs.CL（2025-07-29）

📊 共 21 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) (14 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (14 篇)

#题目一句话要点标签🔗⭐
1 A Scalable Pipeline for Estimating Verb Frame Frequencies Using Large Language Models 提出一种可扩展的流水线，利用大型语言模型估计动词框架频率。 large language model
2 Predicting Microbial Ontology and Pathogen Risk from Environmental Metadata with Large Language Models 利用大语言模型，仅凭环境元数据预测微生物本体和病原体风险 large language model
3 AgriEval: A Comprehensive Chinese Agricultural Benchmark for Large Language Models 提出 AgriEval：首个全面的中文农业大语言模型评测基准 large language model ✅
4 Cyber-Zero: Training Cybersecurity Agents without Runtime Cyber-Zero：无需运行时环境训练网络安全智能体，实现卓越性能。 large language model
5 Meaning-infused grammar: Gradient Acceptability Shapes the Geometric Representations of Constructions in LLMs 研究表明大型语言模型通过梯度可接受性学习并几何表示了语言结构。 large language model
6 Persona-Augmented Benchmarking: Evaluating LLMs Across Diverse Writing Styles 提出Persona增强基准测试，评估LLM在多样化写作风格下的性能。 large language model
7 IndoPref: A Multi-Domain Pairwise Preference Dataset for Indonesian IndoPref：首个印尼语多领域成对偏好数据集，用于评估LLM生成文本的自然性和质量。 large language model
8 Culinary Crossroads: A RAG Framework for Enhancing Diversity in Cross-Cultural Recipe Adaptation 提出CARRIAGE框架，增强RAG在跨文化食谱改编中的多样性，提升用户体验。 large language model
9 Who's important? -- SUnSET: Synergistic Understanding of Stakeholder, Events and Time for Timeline Generation SUnSET：融合利益相关者、事件和时间信息，用于新闻时间线生成，达到SOTA。 large language model
10 The Problem with Safety Classification is not just the Models 揭示多语言安全分类模型及评估数据集的局限性，促进更有效的有害内容识别。 large language model
11 UnsafeChain: Enhancing Reasoning Model Safety via Hard Cases UnsafeChain：通过难例提升推理模型安全性 chain-of-thought ✅
12 TriangleMix: Accelerating Prefilling via Decoding-time Contribution Sparsity TriangleMix：通过解码时贡献稀疏性加速LLM的Prefilling阶段 large language model
13 Persona Vectors: Monitoring and Controlling Character Traits in Language Models 提出Persona Vectors，用于监控和控制语言模型中的人格特质。 large language model
14 VN-MTEB: Vietnamese Massive Text Embedding Benchmark 提出VN-MTEB越南语大规模文本嵌入基准，用于评估和提升越南语NLP模型性能。 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🔗⭐
15 Post-Training Large Language Models via Reinforcement Learning from Self-Feedback 提出基于自反馈强化学习的LLM后训练方法，提升校准性和推理能力 reinforcement learning RLHF large language model
16 RL from Teacher-Model Refinement: Gradual Imitation Learning for Machine Translation 提出RLfR：通过教师模型精炼的强化学习用于机器翻译，提升语义质量和实体保持。 reinforcement learning imitation learning preference learning
17 AutoTIR: Autonomous Tools Integrated Reasoning via Reinforcement Learning AutoTIR：通过强化学习实现自主工具集成推理 reinforcement learning large language model ✅
18 DeepSieve: Information Sieving via LLM-as-a-Knowledge-Router 提出DeepSieve，通过LLM作为知识路由器的信息筛选框架，提升RAG在复杂问答中的性能。 distillation large language model ✅
19 Graph-R1: Towards Agentic GraphRAG Framework via End-to-end Reinforcement Learning 提出Graph-R1框架以解决传统RAG方法的结构语义不足问题 reinforcement learning
20 Libra: Assessing and Improving Reward Model by Learning to Think 提出Libra框架，评估并提升奖励模型在复杂推理场景下的性能。 reinforcement learning large language model
21 Multi-Hypothesis Distillation of Multilingual Neural Translation Models for Low-Resource Languages 提出多假设蒸馏方法，提升低资源语言神经机器翻译模型性能 distillation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页