cs.CL（2025-12-25）

📊 共 14 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (7) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱一：机器人控制 (Robot Control) (2) 支柱六：视频提取与匹配 (Video Extraction) (2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (7 篇)

#题目一句话要点标签🔗⭐
1 Do Latent Tokens Think? A Causal and Adversarial Analysis of Chain-of-Continuous-Thought 揭示潜在令牌的伪推理本质：对Chain-of-Continuous-Thought进行因果和对抗分析 large language model chain-of-thought
2 Enabling Conversational Behavior Reasoning Capabilities in Full-Duplex Speech 提出基于思维图谱的对话行为推理框架，提升全双工语音交互系统的自然性。 multimodal chain-of-thought
3 Five Years of SciCap: What We Learned and Future Directions for Scientific Figure Captioning SciCap项目五年回顾：总结科学图像描述经验，展望未来研究方向 large language model
4 Context Discipline and Performance Correlation: Analyzing LLM Performance and Quality Degradation Under Varying Context Lengths 分析上下文长度变化对LLM性能和质量的影响，揭示KV缓存瓶颈 large language model
5 MoRAgent: Parameter Efficient Agent Tuning with Mixture-of-Roles MoRAgent：基于混合角色（MoR）的参数高效Agent微调框架 large language model
6 Heaven-Sent or Hell-Bent? Benchmarking the Intelligence and Defectiveness of LLM Hallucinations 提出HIC-Bench，用于评估LLM幻觉的智能与缺陷，并探索其在科学创新中的作用。 large language model
7 Beyond Heuristics: A Decision-Theoretic Framework for Agent Memory Management 提出DAM：基于决策论的Agent记忆管理框架，解决启发式方法的局限性 large language model

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

#题目一句话要点标签🔗⭐
8 Compass-Embedding v4: Robust Contrastive Learning for Multilingual E-commerce Embeddings Compass-Embedding v4：面向东南亚电商场景的鲁棒对比学习多语言嵌入 representation learning contrastive learning
9 A Unified Definition of Hallucination: It's The World Model, Stupid! 统一幻觉定义：核心在于语言模型对世界的建模能力 world model
10 Rethinking Sample Polarity in Reinforcement Learning with Verifiable Rewards 提出A3PO：一种自适应非对称token级优势塑造策略优化方法，提升基于可验证奖励的强化学习推理能力。 reinforcement learning

🔬 支柱一：机器人控制 (Robot Control) (2 篇)

#题目一句话要点标签🔗⭐
11 Break Out the Silverware -- Semantic Understanding of Stored Household Items 提出NOAM模型，解决服务机器人居家环境中物品存储位置的语义理解难题 manipulation scene understanding large language model
12 Detecting AI-Generated Paraphrases in Bengali: A Comparative Study of Zero-Shot and Fine-Tuned Transformers 研究孟加拉语AI生成释义文本检测，对比零样本与微调Transformer模型性能。 manipulation large language model

🔬 支柱六：视频提取与匹配 (Video Extraction) (2 篇)

#题目一句话要点标签🔗⭐
13 Oogiri-Master: Benchmarking Humor Understanding via Oogiri Oogiri-Master：通过大喜利游戏基准测试语言模型的幽默理解能力 HuMoR large language model
14 WearVox: An Egocentric Multichannel Voice Assistant Benchmark for Wearables WearVox：面向可穿戴设备的多通道语音助手评测基准 egocentric large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页