cs.CL（2026-03-18）

📊 共 18 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (14 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (2 🔗1) 支柱六：视频提取与匹配 (Video Extraction) (1) 支柱四：生成式动作 (Generative Motion) (1)

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

#题目一句话要点标签🔗⭐
1 Process Supervision for Chain-of-Thought Reasoning via Monte Carlo Net Information Gain 提出基于信息论的过程监督方法以提升推理可靠性 large language model chain-of-thought
2 Inducing Epistemological Humility in Large Language Models: A Targeted SFT Approach to Reducing Hallucination 提出HypoTermInstruct数据集，通过针对性SFT提升LLM的认知谦逊性，减少幻觉 large language model
3 Zipper-LoRA: Dynamic Parameter Decoupling for Speech-LLM based Multilingual Speech Recognition 提出Zipper-LoRA，解决语音LLM多语种语音识别中的稳定性-可塑性困境。 large language model language conditioned ✅
4 TRiMS: Real-Time Tracking of Minimal Sufficient Length for Efficient Reasoning via RL TRiMS：通过强化学习实时追踪最小充分长度，实现高效推理 large language model chain-of-thought
5 IndicSafe: A Benchmark for Evaluating Multilingual LLM Safety in South Asia IndicSafe：评估南亚多语种LLM安全性的基准 large language model
6 Mitigating LLM Hallucinations through Domain-Grounded Tiered Retrieval 提出领域知识增强的分层检索框架，缓解大语言模型幻觉问题 large language model
7 Detecting the Machine: A Comprehensive Benchmark of AI-Generated Text Detectors Across Architectures, Domains, and Adversarial Conditions 构建AI生成文本检测的综合基准，评估多种架构、领域和对抗条件下的检测器性能 large language model
8 Efficient Training-Free Multi-Token Prediction via Embedding-Space Probing 提出基于嵌入空间探测的无训练多Token预测方法，提升LLM推理效率。 large language model
9 DebugLM: Learning Traceable Training Data Provenance for LLMs DebugLM：学习LLM的可追溯训练数据来源，实现行为溯源与修正。 large language model
10 Do Language Models Encode Semantic Relations? Probing and Sparse Feature Analysis 通过探针和稀疏特征分析，研究语言模型对语义关系的编码能力 large language model
11 KA2L: A Knowledge-Aware Active Learning Framework for LLMs 提出KA2L框架，通过知识感知主动学习提升LLM领域知识掌握能力 large language model
12 Language on Demand, Knowledge at Core: Composing LLMs with Encoder-Decoder Translation Models for Extensible Multilinguality 提出XBridge，利用翻译模型增强LLM的多语言能力，解决低资源语言性能瓶颈。 large language model
13 SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems 提出SafeTutors基准以评估AI辅导系统的教学安全性 large language model
14 Beyond bouba/kiki: Multidimensional semantic signals are deeply woven into the fabric of natural language 揭示自然语言中音素与多维语义信号的深刻关联 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (2 篇)

#题目一句话要点标签🔗⭐
15 CoVerRL: Breaking the Consensus Trap in Label-Free Reasoning via Generator-Verifier Co-Evolution CoVerRL：通过生成器-验证器协同进化打破无标签推理中的共识陷阱 reinforcement learning large language model
16 From Isolated Scoring to Collaborative Ranking: A Comparison-Native Framework for LLM-Based Paper Evaluation 提出CNPE框架，将LLM论文评估从孤立评分转向协同排序，提升泛化能力。 reinforcement learning large language model ✅

🔬 支柱六：视频提取与匹配 (Video Extraction) (1 篇)

#题目一句话要点标签🔗⭐
17 Harm or Humor: A Multimodal, Multilingual Benchmark for Overt and Covert Harmful Humor 提出多模态多语言基准以解决有害幽默检测问题 HuMoR multimodal

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
18 Adaptive Guidance for Retrieval-Augmented Masked Diffusion Models 提出ARAM框架，通过自适应指导解决检索增强扩散模型中的检索先验冲突问题。 MDM

⬅️ 返回 cs.CL 首页 · 🏠 返回主页