cs.CL（2026-08-27）

📊 共 36 篇论文 | 🔗 8 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (26 🔗6) 支柱二：RL算法与架构 (RL & Architecture) (10 🔗2)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (26 篇)

#题目一句话要点标签🔗⭐
1 Said Aloud, Read Different: Cross-Modal Instability in Multimodal Models 提出跨模态不稳定性基准以解决多模态模型一致性问题 foundation model multimodal
2 Mapping Written Words to Spoken Words in a Different Language Using Only Visual Grounding 提出基于视觉对齐的方法以解决跨语言语音映射问题 multimodal visual grounding
3 Dependency-Aware Revocable Decoding for Efficient Diffusion Large Language Model Inference 提出依赖感知可撤销解码以提升扩散大语言模型推理效率 large language model multimodal
4 SCIT: Testing Causal Cache Carriers in Latent Chain-of-Thought Models 提出SCIT以解决潜在链式思维模型中的因果缓存问题 chain-of-thought
5 Prediction of Prediction (PoP): Inter-Layer Activation Fusion for Single-Pass Hallucination Detection in Large Language Models 提出PoP机制以解决大型语言模型的事实错误检测问题 large language model
6 RuleWeaver: Benchmarking Rule-Centered Scenario Reasoning for Large Language Models 提出RuleWeaver以解决大型语言模型的规则中心场景推理问题 large language model ✅
7 Benchmarking Clinical Decision Pathway Adherence in Large Language Models 提出MEGA-CDP以评估医疗大语言模型的临床决策路径遵循性 large language model
8 Surgical Alignment in Knowledge Graph Training for Clinical Diagnosis with Large Language Models 提出KG信号整合方法以提升临床诊断中的LLM推理能力 large language model ✅
9 Double Trouble: Bilingual Pretraining Leaves Language-Conditioned Effects in Shared-Language Representations 揭示双语预训练模型共享语言表示的隐藏状态差异 language conditioned
10 RCMN: Understanding Misleadingness in Influential Public Discourse 提出RCMN框架以理解公共话语中的误导性问题 foundation model multimodal
11 INTENT-AS-A-TOOL Makes it Easy to Track Agentic Misalignment 提出INTENT-AS-A-TOOL以解决代理性不一致问题 large language model chain-of-thought ✅
12 DocTalkBN: A Novel Dataset of Expert Telemedicine Conversations in Bengali 提出DocTalkBN数据集以解决孟加拉语医疗对话数据匮乏问题 large language model multimodal
13 Reasoning about In-Context Samples for Machine-Translation 提出基于片段推理的机器翻译方法以提升翻译可靠性 large language model chain-of-thought
14 How Language Models Organize and Structure Moral Knowledge 提出线性探针以探讨语言模型中的道德知识结构 large language model
15 Do LLMs Understand Personality? Rethinking Persona Fidelity Evaluation through Structured Behavioral Inference 提出PRISM框架以解决大语言模型的人格一致性评估问题 large language model
16 CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes 提出CritICL框架以提升小型语言模型推理效率 large language model ✅
17 BALMS: Benchmarking Agentic LLMs for Longitudinal Mental Health Sensing 提出BALMS基准以解决长期心理健康感知问题 chain-of-thought
18 JudgeStealer: Extracting LLM Judging Capabilities across Evaluation Protocols 提出JudgeStealer以高效提取LLM评判能力 large language model
19 Information-Guided Frontier Decoding: Contextual Utility-Driven Commitment in dMLLMs 提出信息引导前沿解码以提升多模态语言模型的解码质量 multimodal
20 Sycophancy Suppression Can Impair Rational Updating: Anti-Sycophancy Should Preserve the Ability to Update 提出反谄媚策略以平衡理性更新与支持性妥协问题 large language model
21 When Tokenizers Fail: Byte-Level Chunking for Zero-Shot Transfer to Low-Resource Languages 提出适应性层次网络框架以解决低资源语言处理问题 zero-shot transfer
22 Surgical Alignment in Knowledge Graph Training for Clinical Diagnosis with Large Language Models 提出KG信号整合方法以提升临床诊断中的LLM推理能力 large language model ✅
23 Memorization Is Not Extraction: Tight Differential-Privacy Bounds and Audit Blind Spots 提出精确的差分隐私界限以解决记忆与提取问题 large language model
24 Compositional Failure in Audio-Visual LLMs: Late-Layer Prior Dominance Under Cross-modal Conflict 研究音频-视觉LLM中的组合失败问题，提出新分析方法 instruction following ✅
25 Load-Bearing Context: The Question Damage Score for Evaluating Context Reliance in Linguistic Reasoning 提出问题损伤评分以评估语言推理中的上下文依赖性 large language model
26 INSPIRE: An Internalize-Then-Improve Approach for Example-Driven Mathematical Reasoning 提出INSPIRE以解决数学推理模型内化不足问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (10 篇)

#题目一句话要点标签🔗⭐
27 FOCUS & RePAIR: Mitigating Text Degeneration via Token-Level Guidance for Pruned Large Language Models 提出FOCUS与RePAIR以缓解修剪大语言模型的文本退化问题 distillation large language model
28 Not Just Reason, Not Just Scan: Reinforcement Learning for Proactive Scientific Error Verification over Academic Paper 提出VERA-RL以解决学术论文中的科学错误验证问题 reinforcement learning large language model multimodal
29 Preserving General Capabilities during Domain Specialization with Uncertainty-Calibrated MOPD 提出不确定性校准的多教师在线蒸馏以解决领域专门化问题 distillation large language model instruction following
30 TTPO: Test-Time Policy Optimization 提出TTPO以解决测试时训练中的标签依赖问题 reinforcement learning distillation large language model
31 Consolidating RLVR Capabilities Across Domains: A Deep Dive into Fusion Paradigms 提出三种融合范式以优化多领域强化学习能力 reinforcement learning distillation large language model
32 SPEAR: Distilling Domain-Adaptive Reasoning Skeletons via Sequential Symbolic Alignment in Reinforcement Learning 提出SPEAR以解决强化学习中的推理蒸馏问题 reinforcement learning distillation ✅
33 Instruction Quality Matters: Refining Instructions for Effective Preference Learning 提出指令优化管道以提升偏好学习效果 preference learning ✅
34 Boosting LLM Exploration via Weak-Model Guidance in RLVR 通过弱模型引导提升RLVR中的LLM探索能力 reinforcement learning reward design
35 A Survey on Rubric-Guided Reinforcement Learning for Language Models 提出基于评分标准的强化学习以解决语言模型对人类偏好的对齐问题 reinforcement learning RLHF reward design
36 Below the Noise Floor: Bimodal Seed Collapse and Distinct Failure Modes in Small-Model Knowledge Distillation 提出双模种子崩溃机制以解决小模型知识蒸馏中的失败模式问题 distillation

⬅️ 返回 cs.CL 首页 · 🏠 返回主页