cs.CL（2026-01-12）

📊 共 31 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) (24 🔗2) 支柱二：RL算法与架构 (RL & Architecture) (6) 支柱一：机器人控制 (Robot Control) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (24 篇)

#题目一句话要点标签🔗⭐
1 A Unified Framework for Emotion Recognition and Sentiment Analysis via Expert-Guided Multimodal Fusion with Large Language Models 提出EGMF框架，通过专家引导的多模态融合与大语言模型，统一解决情感识别和情感分析任务。 large language model multimodal
2 PlaM: Training-Free Plateau-Guided Model Merging for Better Visual Grounding in MLLMs 提出PlaM，通过无训练的平台引导模型融合，提升MLLM的视觉定位能力。 large language model multimodal visual grounding ✅
3 Enhancing Self-Correction in Large Language Models through Multi-Perspective Reflection 提出PR-CoT，通过多视角反思增强大语言模型的自我纠错能力 large language model chain-of-thought
4 BayesRAG: Probabilistic Mutual Evidence Corroboration for Multimodal Retrieval-Augmented Generation 提出BayesRAG，通过概率互证提升多模态检索增强生成效果 large language model multimodal ✅
5 Thinking Before Constraining: A Unified Decoding Framework for Large Language Models 提出一种统一解码框架，结合自然语言推理和结构化生成，提升大语言模型性能。 large language model
6 Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models 提出Engram，通过可扩展查找的条件记忆，为大语言模型引入新的稀疏性维度。 large language model
7 Can Large Language Models Understand, Reason About, and Generate Code-Switched Text? 提出CodeMixQA基准，评估LLM在代码切换文本理解、推理和生成能力 large language model
8 ReMIND: Orchestrating Modular Large Language Models for Controllable Serendipity A REM-Inspired System Design for Emergent Creative Ideation ReMIND：一种用于可控偶然发现的模块化大语言模型编排系统，用于涌现式创意构思 large language model
9 Structure First, Reason Next: Enhancing a Large Language Model using Knowledge Graph for Numerical Reasoning in Financial Documents 提出一种基于知识图谱增强的大语言模型，用于金融文档中的数值推理。 large language model
10 Exploring the Meta-level Reasoning of Large Language Models via a Tool-based Multi-hop Tabular Question Answering Task 提出基于工具的多跳表格问答任务，探索大语言模型的元层次推理能力 large language model
11 DiffER: Diffusion Entity-Relation Modeling for Reversal Curse in Diffusion Large Language Models 提出DiffER以解决扩散大语言模型中的反转诅咒问题 large language model
12 ActiShade: Activating Overshadowed Knowledge to Guide Multi-Hop Reasoning in Large Language Models ActiShade：激活被掩盖知识，引导大语言模型进行多跳推理 large language model
13 MI-PRUN: Optimize Large Language Model Pruning via Mutual Information 提出MI-PRUN：通过互信息优化大语言模型剪枝 large language model
14 The Confidence Trap: Gender Bias and Predictive Certainty in LLMs 提出Gender-ECE指标，评估LLM在性别偏见下的置信度校准问题 large language model
15 Adaptive Layer Selection for Layer-Wise Token Pruning in LLM Inference 提出ASL自适应选择LLM推理中逐层token剪枝的层，提升KV缓存效率。 large language model
16 Kinship Data Benchmark for Multi-hop Reasoning 提出KinshipQA基准，用于评估LLM在亲属关系推理中的多跳推理能力。 large language model
17 Proof of Time: A Benchmark for Evaluating Scientific Idea Judgments 提出PoT：一个评估科学研究想法判断质量的半可验证基准框架。 large language model
18 Two Pathways to Truthfulness: On the Intrinsic Encoding of LLM Hallucinations 揭示LLM幻觉的内在编码机制：问题锚定与答案锚定双路径 large language model
19 Learning Through Dialogue: Unpacking the Dynamics of Human-LLM Conversations on Political Issues 通过对话学习：剖析人-LLM就政治议题对话的动态机制 large language model
20 Is Agentic RAG worth it? An experimental comparison of RAG approaches 对比增强型与Agentic RAG：探究不同RAG方法在实际应用中的权衡 large language model
21 High-Rank Structured Modulation for Parameter-Efficient Fine-Tuning 提出SMoA：一种高秩结构化调制适配器，用于参数高效的微调。 large language model
22 On Narrative: The Rhetorical Mechanisms of Online Polarisation 提出叙事极化度量方法，分析在线冲突中叙事如何塑造群体极化。 large language model
23 GROKE: Vision-Free Navigation Instruction Evaluation via Graph Reasoning on OpenStreetMap GROKE：基于OpenStreetMap图推理的免视觉导航指令评估框架 VLN
24 Document-Level Zero-Shot Relation Extraction with Entity Side Information 提出DocZSRE-SI框架，利用实体侧信息解决文档级零样本关系抽取问题，提升低资源语言性能。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (6 篇)

#题目一句话要点标签🔗⭐
25 Controlling Multimodal Conversational Agents with Coverage-Enhanced Latent Actions 提出覆盖增强的潜在动作方法，用于控制多模态对话Agent。 reinforcement learning multimodal
26 Structured Reasoning for Large Language Models 提出结构化推理(SCR)框架，提升大语言模型推理效率和自验证能力。 reinforcement learning large language model
27 The Confidence Dichotomy: Analyzing and Mitigating Miscalibration in Tool-Use Agents 提出基于强化学习的微调框架，提升工具使用Agent的校准性并解决置信度二分问题。 reinforcement learning reward design large language model
28 ReinPool: Reinforcement Learning Pooling Multi-Vector Embeddings for Retrieval System 提出ReinPool，利用强化学习池化多向量嵌入，显著压缩索引并提升检索性能。 reinforcement learning
29 GRPO with State Mutations: Improving LLM-Based Hardware Test Plan Generation 提出GRPO-SMu，提升LLM在硬件测试计划生成中的性能 reinforcement learning large language model
30 Contrastive Learning with Narrative Twins for Modeling Story Salience 提出基于叙事孪生的对比学习框架，用于建模故事显著性。 contrastive learning

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
31 KALE: Enhancing Knowledge Manipulation in Large Language Models via Knowledge-aware Learning KALE：通过知识感知学习增强大语言模型中的知识操控能力 manipulation large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页