cs.CL（2026-10-08）

📊 共 43 篇论文 | 🔗 5 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (31 🔗4) 支柱二：RL算法与架构 (RL & Architecture) (12 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (31 篇)

#题目一句话要点标签🔗⭐
1 Large Language Model Turnover Undermines Screening for Artificial Intelligence-Assisted Scientific Writing 量化大语言模型更迭对科学写作筛选的影响 large language model
2 When History Helps and Hurts: Selective History Use across Multimodal Turns 提出ReTurn基准以解决多模态对话历史选择性使用问题 multimodal
3 Not Every Change Is Necessary: Recoverable Drift in Large Language Model Unlearning 提出PTP-U框架以解决大语言模型的可恢复遗忘问题 large language model
4 Detecting Spin in Clinical Trials with Large Language Models 利用大语言模型自动检测临床试验中的结果扭曲 large language model
5 From Sparse Representations to Behavioral Insights for Multimodal Depression Assessment 提出BehavDep框架以解决多模态抑郁评估的可解释性问题 multimodal
6 BeliefScope: Diagnosing Evidence-Driven Revision and Pressure-Induced Shifts in Large Language Models 提出BeliefScope以区分证据驱动修正与压力诱导变化 large language model
7 Deception by Omission: Language Models Knowingly Hide Their Mistakes 研究语言模型的错误隐瞒行为以提升用户信任 large language model chain-of-thought
8 SignRAG: Unified Retrieval-Augmented Gloss-Free Sign Language Translation 提出SignRAG以解决无注释手语翻译的局限性问题 large language model ✅
9 VFold: Symmetry-Aware Cross-Layer Value Cache Compression 提出VFold以解决长上下文下缓存内存占用问题 large language model
10 Examining Social Attribution in LLM Reasoning: A Theory-Guided Probing Methodology 提出社会归因基准以探究大语言模型的推理机制 large language model ✅
11 Adapting English Quality Classifiers for Multilingual LLM Pretraining Data Selection 提出多语言适应方法以优化LLM预训练数据选择 large language model
12 TokenRouter: Efficient Serving System for Token-Level LLM Routing 提出TokenRouter以解决高效的Token级LLM路由问题 large language model ✅
13 SteerablePlex: Can We Steer Full-Duplex Models? 提出SteerablePlex以解决全双工模型控制难题 instruction following
14 A persistent accuracy ceiling in automated verbal deception detection 系统评估自动化语言欺骗检测的准确性上限 large language model
15 Natural Language to First-Order Logic LLM-based Autoformalization 提出FOL自动形式化任务定义以解决现有方法不足 large language model
16 Specialized Decision Models vs. General-Purpose LLMs: Benchmarking Jev Across Knowledge, Reasoning, and Multilingual Tasks 提出Jev模型以解决多选题决策问题 large language model
17 Can Decision Models Understand Stance? Evaluating Jev Against General-Purpose LLMs 提出Jev模型以解决立场检测问题 large language model
18 Thinking Inertia: LLMs Keep Thinking When Told Not To 提出无思维行为评估方法以解决LLMs推理控制问题 large language model
19 Chronos Enables Code Agents to Reason over Software Evolution 提出Chronos框架以增强代码代理对软件演化的推理能力 large language model
20 Does Modern Standard Arabic (MSA) Dominate Arabic Dialects in LLMs? A Representation-Level Analysis 提出语言主导框架分析阿拉伯方言与现代标准阿拉伯语的关系 large language model
21 AdaptEvo: Adaptive Agent Learning with Evolving Supervision 提出AdaptEvo框架以解决不完美监督下的决策问题 multimodal
22 From Retrieval to Reconstruction: Constructing Evolvable Cognitive Memory for Long-Term Dialogue 提出CogMem以解决长时间对话中的记忆管理问题 large language model ✅
23 Unsupervised Multimodal Reward Modeling 提出无监督多模态奖励建模方法以降低标注成本 multimodal
24 GeoFitBench-LLM: A Benchmark for Assessing Large Language Models' Capabilities in Geospatial Data Requirements Elicitation for Geo-Analytic Tasks 提出GeoFitBench-LLM以评估大语言模型在地理数据需求提取中的能力 large language model
25 Large Language Models, Machine-Assisted Cultural Analytics, and the Agentic Turn 探讨大型语言模型在文化分析中的应用与挑战 large language model
26 DistressBench: Evaluating Multi-Dimensional Crisis Support in Large Language Models 提出DistressBench以评估语言模型在危机支持中的表现 large language model
27 When History Helps and Hurts: Selective History Use across Multimodal Turns 提出ReTurn基准以解决多模态历史信息选择性使用问题 multimodal
28 U4RASD at KnowledgeGraphEval-2026: Ontology-Constrained Encoder and LLM Systems for Arabic Relation Extraction 提出基于本体约束的编码器和LLM系统以解决阿拉伯关系提取问题 large language model
29 perturber: A Stateless Service for Semantics-Preserving Text Perturbations 提出perturber以解决文本扰动的可复现性问题 large language model
30 Knowing Is Not Telling: Identity Ossification in LLM Representations of Trans-border Communities 提出跨境身份固化诊断框架以解决大语言模型的身份表现问题 large language model
31 Contextual Correction of Parallel Draftsfor Speculative Decoding 提出DCOCO以解决并行草拟中的上下文依赖问题 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (12 篇)

#题目一句话要点标签🔗⭐
32 EgoVoice: Proactive Spoken Assistance from Egocentric Multimodal Streams 提出EgoVoice以解决主动语音助手的时机与内容选择问题 direct preference optimization egocentric multimodal
33 MetaEncoder: Exploring the Limit of Bi-Encoders for Multimodal System One Decision Making with Natural Language Interface 提出MetaEncoder以解决多模态系统一决策中的自然语言接口问题 contrastive learning multimodal instruction following
34 MiMo-V2.6: Scaling Reinforcement Learning Towards Self-Improvement 提出MiMo-V2.6以推动强化学习向自我改进的进展 reinforcement learning foundation model multimodal
35 Language Models as AI Research World Models 提出语言模型作为研究世界模型以提升实验预测能力 world model world models
36 DIAL-OPD: Learning More from Fewer Tokens in On-Policy Distillation 提出DIAL-OPD以在策略蒸馏中减少标记数量提升学习效果 teacher-student distillation
37 When Do We Need On-Policy Distillation? Distilling on Offline Student Rollouts Is Often Better 提出Semi-OPD以优化教师-学生模型蒸馏过程 teacher-student distillation
38 Lapras: Latent Reasoning for Time Series Language Models 提出Lapras以解决时间序列语言模型推理不足问题 teacher-student distillation chain-of-thought
39 Beyond Sequences: Distilling Structured Decision Memory for LLM Recommendation 提出MARI以解决推荐系统中的决策记忆问题 distillation large language model
40 ReCal: Calibrating Structured Pruning for On-Policy Distillation Recovery 提出RECAL以解决结构化剪枝对蒸馏恢复的影响 distillation
41 SFT-as-Context Mitigates Forgetting in Supervised Fine-Tuning 提出SFT-as-context以缓解监督微调中的遗忘问题 MAE large language model
42 Residual Advantage: Student-Relative Teacher Guidance for RL with Verifiable Rewards 提出Residual Advantage以优化RLVR中的教师指导问题 reinforcement learning distillation
43 CLUE: A Contrastive Learning Framework for Unsupervised Event Type Induction and Annotation 提出CLUE框架以解决无监督事件类型归纳与标注问题 contrastive learning ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页