cs.CL（2026-10-13）

📊 共 25 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (18 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗3)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (18 篇)

#题目一句话要点标签🔗⭐
1 The Image Still Has a Say: Final-Answer Repair for Multimodal Chain-of-Thought 提出最终答案修复方法以解决多模态推理中的答案准确性问题 multimodal chain-of-thought
2 Code-Switching Detection via Language-Selective Dimensions in Speech Foundation Models 提出一种轻量级方法直接从语音中检测代码切换 foundation model
3 From Hindsight to Foresight: Dissecting Clinical Evidence Acquisition in Large Language Models 提出从事后分析到前瞻性诊断的方法以解决LLMs的临床证据获取问题 large language model
4 Are We Testing Too Much? A Survey of Benchmark Compression for Efficient Evaluation of Large Language Models 提出基准压缩方法以降低大型语言模型评估成本 large language model
5 Beyond Belief Nesting: Rethinking Higher-Order Theory of Mind in Large Language Models 提出HopToM与HopTrace以解决高阶心智理论评估问题 large language model
6 Understanding Machine Taste: Aesthetic Agreement Across Foundation Models in K-pop 提出机器审美理解方法以评估K-pop创意内容 foundation model
7 Decouple to Forget: Semantic Mediation for Sensitive Unlearning in Large Language Models 提出D2F方法以解决大语言模型中的敏感信息遗忘问题 large language model ✅
8 From Relation-Specific Heads to Shared Interventions: Contextual Entrainment in Language Models 提出共享干预机制以解决语言模型的上下文依赖问题 large language model
9 Jev-ToM: Evaluating Theory of Mind in Typed Decision Models 提出Jev-ToM以优化理论心智模型的选择过程 large language model
10 The Many Morals of AI: The cautionary bias of North American news narratives in global perspective 分析北美新闻叙事对人工智能的警示性偏见 large language model
11 Lightweight but Linguistically Fair? Quantisation Robustness across Bavarian, German, and English 研究量化鲁棒性以解决区域语言模型的公平性问题 large language model
12 Virtual Head Attention 提出虚拟头注意力机制以解决长上下文模型的缓存与计算问题 large language model
13 TRACESTEER: Trace Driven Activation Steering for Hallucination Mitigation in Diffusion Language Models 提出TRACESTEER以解决扩散语言模型中的幻觉问题 large language model
14 SparPR: Runtime KV-Cache Recovery for Resource-Constrained LLM Inference on Consumer Hardware 提出SparPR以解决消费硬件上LLM推理的KV缓存问题 large language model
15 Machine--human agreement in marking undergraduate assignments is specific to model, prompt and institution 评估本科作业的机器与人类评分一致性受模型、提示和机构影响 large language model
16 Perturbation, Not Provenance: Why Activation Steering Does Not Make a Robust Text Watermark 提出激活引导方法以解决文本水印鲁棒性不足问题 large language model
17 GENDIS: Gender-Related Discourse and Stance in Multilingual YouTube Comments from India, Kenya, and Nigeria 提出GENDIS数据集以解决性别相关在线内容研究不足问题 large language model
18 Encoding Context into LLM Weights: What Makes It Context? 提出上下文编码方法以提升大语言模型的推理能力 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🔗⭐
19 Do Factuality Methods Make LLMs More Factual? Reproducibility and Fair Comparison 首次独立复现五种事实性对齐方法以评估其有效性 PPO RLHF DPO
20 Facts Don't Speak Louder than Words: The Behavioral Essence of Long-CoT Distillation 提出行为对齐理论以优化长链思维蒸馏方法 distillation chain-of-thought ✅
21 Imprinting: Knowledge Distillation Can Shift Student Behavior 提出印记机制以解决知识蒸馏中的行为偏差问题 teacher-student distillation
22 Agent-Reflex: Teaching Language Agents to Act Reflectively 提出Agent-Reflex以解决语言代理反思学习不足问题 reinforcement learning distillation large language model ✅
23 TextualRL: Textual Reinforcement Learning for Context Management 提出TextualRL以解决文本上下文管理中的奖励转化问题 reinforcement learning
24 Rubrics from the Task, Verdicts from Evidence: Pointwise Rubric Rewards with Tool-Augmented Judges 提出ReVeR以解决现有评估模型的偏差问题 distillation instruction following
25 Fish Audio Drama 3 Technical Report 提出Fish Audio Drama 3以实现高效的文本转语音表演 reinforcement learning linear attention ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页