cs.CL（2026-10-21）

📊 共 27 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) (19 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗2) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (19 篇)

#题目一句话要点标签🔗⭐
1 Omnilingual-MsASR: Benchmarking Multimodal, Multilingual, Long-Form Multi-Speaker ASR of Omni Large Language Models 提出Omnilingual-MsASR以解决多语种长视频多说话人ASR问题 large language model foundation model multimodal ✅
2 Mitigating Query Shortcuts in Chain-of-Thought with Context-Aware Decoding 提出上下文感知解码以解决链式思维中的查询快捷问题 large language model chain-of-thought
3 A Large-Scale Evaluation of Role-Enhanced Zero-Shot Chain-of-Thought Prompting 提出角色增强的零-shot思维链提示框架以提升推理能力 large language model chain-of-thought
4 Benchmarking Large Language Models for Detecting Satirical Code-Mixed Sentences 提出合成代码混合句子的方法以提升讽刺检测能力 large language model chain-of-thought
5 Remembering Through Change: Provenance-Preserving Hypergraph Memory for Multimodal Agents 提出MHG-Mem以解决多模态代理的长期记忆问题 multimodal
6 Investigating the Impact of Learner Errors on Social Bias in Large Language Models 研究非母语者语法错误对大语言模型社会偏见的影响 large language model
7 OmniSimulator: Aligning Small Language Models for Authentic Heterogeneous Behavior Modeling 提出OmniSimulator以解决小型语言模型用户模拟能力不足的问题 large language model chain-of-thought
8 Where Do LLMs Learn to Recommend AI? 探讨大型语言模型推荐AI的来源及影响因素 large language model
9 Every Part Has Its Role: Part-of-Speech Effects in LLM Post-Training 提出基于词性影响的概念增强方法以提升LLM性能 large language model
10 S-EDL: Eliciting Self-Evidence from Sequence Likelihoods for Semantic Calibration of LLMs 提出S-EDL以解决大语言模型语义校准问题 large language model ✅
11 Retrieval Augmented Generation Compositionality 提出LASAGNA框架以优化检索增强生成模型的评估 large language model
12 Memory-Centric Continual Learning for AI Agents 提出记忆中心持续学习方法以解决自主智能体的学习瓶颈问题 foundation model
13 Beyond Binary Labels: Diagnostic Evaluation of LLM Safety with Prompt-Level Rubrics 提出DiagSafe框架以解决LLM安全评估的不足问题 instruction following
14 ALIAS: Evaluating Performance and Memorization in LLM-Written TLA+ Specifications 提出ALIAS方法评估LLM在TLA+规范中的表现与记忆化问题 large language model
15 Accept or Repair: Preserving Passage-Supported Answers in Contrastive Decoding 提出Accept-or-Repair以解决语言模型回答准确性问题 large language model ✅
16 LLMs as survey respondents: Comparison to simple semantic baselines 利用大型语言模型提升问卷调查响应的准确性 large language model
17 Beyond the Vote: When Do LLMs Add Value to Retrieval-Based Tabular Classification? 比较LLM与距离加权KNN在表格分类中的价值 large language model
18 TenetBench: A Clause-Anchored Russian Corpus for Evaluating Language Models against Professional Codes of Ethics 提出TenetBench以评估语言模型对职业伦理规范的遵守情况 large language model
19 DynaWeave: Learning to Adapt through Dynamic Scenario Synthesis 提出DynaWeave以解决动态环境中任务适应性不足的问题 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🔗⭐
20 Right On Q: Recovering Reasoning under Quantization via Self-Generated Trajectories 提出OnQ以解决低比特量化模型推理能力不足问题 teacher-student distillation large language model
21 LongZero: Enhancing Long-Context Reasoning in LLMs from Zero Data 提出LongZero以解决长上下文推理数据稀缺问题 reinforcement learning large language model
22 pplx-embed-v2-late Technical Report 提出pplx-embed-v2-late以解决多模态检索问题 distillation multimodal ✅
23 Target Regeneration Shrinks the Gains of Draft-Training Objectives in Speculative Decoding 提出目标再生方法以缩小草稿训练目标在推测解码中的增益 distillation large language model ✅
24 RCSD: Reinforcement and Continual Learning for Domain-Specific Speculative Decoding 提出RCSD框架以提升领域特定的推测解码效率 distillation large language model
25 Reasoning-Search-Augmented LLMs: An Operational Taxonomy, Benchmark Analysis and Controlled Evidence Synthesis 提出基于推理搜索增强的LLMs以解决信息支持不足问题 reinforcement learning large language model
26 PACE: Periodic and Channel-Aligned Encoding of Four-Bit Recurrent States 提出PACE以解决四位递归状态量化精度不足问题 Mamba linear attention

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
27 Beyond Small Talk: Modeling Disclosure Trajectories in Human-AI Conversation 提出TraDE框架以解决人机对话中的披露轨迹问题 penetration

⬅️ 返回 cs.CL 首页 · 🏠 返回主页