cs.CL（2025-04-14）

📊 共 30 篇论文 | 🔗 6 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (20 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (8 🔗2) 支柱一：机器人控制 (Robot Control) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (20 篇)

#题目一句话要点标签🔗⭐
1 Performance of Large Language Models in Supporting Medical Diagnosis and Treatment 评估大型语言模型在医疗诊断和治疗中的表现，并分析其成本效益。 large language model chain-of-thought
2 VisualPuzzles: Decoupling Multimodal Reasoning Evaluation from Domain Knowledge VisualPuzzles：解耦多模态推理评估与领域知识，专注通用推理能力 large language model multimodal
3 LLM-SRBench: A New Benchmark for Scientific Equation Discovery with Large Language Models 提出LLM-SRBench以解决科学方程发现评估问题 large language model
4 You've Changed: Detecting Modification of Black-Box Large Language Models 提出一种基于文本特征分布比较的黑盒大语言模型修改检测方法 large language model
5 Beyond Chains of Thought: Benchmarking Latent-Space Reasoning Abilities in Large Language Models 提出潜空间推理基准，评估大语言模型在隐式推理中的能力。 large language model
6 Unchecked and Overlooked: Addressing the Checkbox Blind Spot in Large Language Models with CheckboxQA CheckboxQA数据集：解决大语言模型在复选框理解上的盲点 large language model ✅
7 Probing then Editing Response Personality of Large Language Models 提出一种探查与编辑框架，用于控制大型语言模型的回应人格。 large language model ✅
8 The Mirage of Performance Gains: Why Contrastive Decoding Fails to Mitigate Object Hallucinations in MLLMs? 揭示对比解码的局限性：无法有效缓解多模态大语言模型中的对象幻觉问题 large language model multimodal
9 Weight-of-Thought Reasoning: Exploring Neural Network Weights for Enhanced LLM Reasoning 提出Weight-of-Thought推理，利用神经网络权重增强LLM推理能力 large language model chain-of-thought
10 HELIOS: Adaptive Model And Early-Exit Selection for Efficient LLM Inference Serving HELIOS：面向高效LLM推理服务的自适应模型与提前退出选择框架 large language model
11 LLM Unlearning Reveals a Stronger-Than-Expected Coreset Effect in Current Benchmarks LLM卸载研究揭示：现有基准测试中存在超乎预期的 Coreset 效应 large language model ✅
12 Learning from Reference Answers: Versatile Language Model Alignment without Binary Human Preference Data RefAlign：无需二元人类偏好数据，利用参考答案对齐语言模型 large language model ✅
13 HalluSearch at SemEval-2025 Task 3: A Search-Enhanced RAG Pipeline for Hallucination Detection HalluSearch：一种搜索增强的RAG流水线，用于幻觉检测 large language model
14 C-FAITH: A Chinese Fine-Grained Benchmark for Automated Hallucination Evaluation 提出C-FAITH：一个中文细粒度幻觉评估基准，用于自动化评估大语言模型的幻觉问题。 large language model
15 Enhancing Reasoning Abilities of Small LLMs with Cognitive Alignment 提出CRV+CogPO框架，提升小模型在复杂推理任务中的认知对齐能力 chain-of-thought
16 DioR: Adaptive Cognitive Detection and Contextual Retrieval Optimization for Dynamic Retrieval-Augmented Generation 提出DioR，自适应认知检测与上下文检索优化动态RAG，提升LLM生成质量。 large language model
17 DataPuzzle: Breaking Free from the Hallucinated Promise of LLMs in Data Analysis DataPuzzle：提出多智能体框架，解决LLM在数据分析中幻觉问题，提升可信度。 large language model
18 C-MTCSD: A Chinese Multi-Turn Conversational Stance Detection Dataset 提出C-MTCSD：一个大规模中文多轮对话立场检测数据集，用于提升社交媒体分析。 large language model
19 Guiding Reasoning in Small Language Models with LLM Assistance SMART框架：利用LLM辅助小模型进行复杂推理 large language model
20 Augmented Relevance Datasets with Fine-Tuned Small LLMs 利用微调的小型LLM增强相关性数据集，提升排序模型性能 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (8 篇)

#题目一句话要点标签🔗⭐
21 Localized Cultural Knowledge is Conserved and Controllable in Large Language Models 提出文化定制向量，提升大语言模型在多语言场景下的文化适应性和可控性。 world model large language model
22 Learning to Erase Private Knowledge from Multi-Documents for Retrieval-Augmented Large Language Models 提出Eraser4RAG，用于检索增强大语言模型中多文档的隐私知识擦除。 PPO large language model
23 MT-R1-Zero: Advancing LLM-based Machine Translation via R1-Zero-like Reinforcement Learning 提出MT-R1-Zero，通过类R1-Zero强化学习提升LLM机器翻译能力，无需监督微调或冷启动。 reinforcement learning reward design large language model ✅
24 SocioVerse: A World Model for Social Simulation Powered by LLM Agents and A Pool of 10 Million Real-World Users SocioVerse：基于LLM Agent和千万级真实用户池的社会模拟世界模型 world model large language model
25 Transferable text data distillation by trajectory matching 提出基于轨迹匹配的可迁移文本数据蒸馏方法，降低大语言模型训练成本。 distillation large language model
26 Improving In-Context Learning with Reasoning Distillation 提出推理蒸馏技术以提升语言模型的上下文学习能力 imitation learning distillation ✅
27 Better Estimation of the Kullback--Leibler Divergence Between Language Models 提出Rao-Blackwellized估计器，更稳定地估计语言模型间的KL散度。 reinforcement learning RLHF distillation
28 DeepTrans: Deep Reasoning Translation via Reinforcement Learning 提出DeepTrans，通过强化学习实现深度推理的自由翻译 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
29 LLM Can be a Dangerous Persuader: Empirical Study of Persuasion Safety in Large Language Models 提出PersuSafety框架，系统评估大型语言模型在劝说场景中的安全性风险。 manipulation large language model
30 Characterizing Knowledge Manipulation in a Russian Wikipedia Fork 提出方法分析俄罗斯维基百科分支中的知识操控现象 manipulation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页