cs.CL（2026-01-09）

📊 共 32 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) (25 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (7 🔗3)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (25 篇)

#题目一句话要点标签🔗⭐
1 The Molecular Structure of Thought: Mapping the Topology of Long Chain-of-Thought Reasoning 提出Mole-Syn方法，通过合成有效长链思考结构，提升LLM推理性能和训练稳定性。 large language model chain-of-thought
2 Multimodal In-context Learning for ASR of Low-resource Languages 提出多模态上下文学习方法，提升低资源语言语音识别性能 large language model multimodal
3 iReasoner: Trajectory-Aware Intrinsic Reasoning Supervision for Self-Evolving Large Multimodal Models iReasoner：轨迹感知的内在推理监督，用于自进化大型多模态模型 multimodal chain-of-thought
4 Afri-MCQA: Multimodal Cultural Question Answering for African Languages 提出Afri-MCQA：一个面向非洲语言的多模态文化问答基准 large language model multimodal ✅
5 Router-Suggest: Dynamic Routing for Multimodal Auto-Completion in Visually-Grounded Dialogs 提出Router-Suggest，用于视觉对话中多模态自动补全的动态路由。 multimodal
6 Continual-learning for Modelling Low-Resource Languages from Large Language Models 提出基于词性编码切换和回放适配器的持续学习方法，缓解低资源语言模型中的灾难性遗忘问题。 large language model
7 Can large language models interpret unstructured chat data on dynamic group decision-making processes? Evidence on joint destination choice 利用大型语言模型解析非结构化聊天数据，辅助动态群体决策过程分析。 large language model
8 Can Large Language Models Differentiate Harmful from Argumentative Essays? Steps Toward Ethical Essay Scoring 提出有害 निबंध 检测基准，揭示大语言模型在识别有害内容方面的不足 large language model
9 Towards Valid Student Simulation with Large Language Models 提出基于大语言模型的学生模拟框架，解决能力悖论问题，提升教育场景的模拟有效性。 large language model
10 Continual Pretraining on Encrypted Synthetic Data for Privacy-Preserving LLMs 提出基于加密合成数据的持续预训练框架，用于保护隐私的大语言模型。 large language model instruction following ✅
11 Simplify-This: A Comparative Analysis of Prompt-Based and Fine-Tuned LLMs Simplify-This：对比Prompt工程与微调LLM在文本简化任务中的性能差异 large language model
12 Visualising Information Flow in Word Embeddings with Diffusion Tensor Imaging 提出基于扩散张量成像的词嵌入信息流可视化方法，提升语言模型可解释性。 large language model
13 Multilingual Amnesia: On the Transferability of Unlearning in Multilingual LLMs 研究多语言LLM中的遗忘学习，揭示跨语言知识迁移对遗忘的影响 large language model
14 Gender Bias in LLMs: Preliminary Evidence from Shared Parenting Scenario in Czech Family Law 评估大型语言模型在捷克家庭法共享育儿场景中的性别偏见 large language model
15 Analysing Differences in Persuasive Language in LLM-Generated Text: Uncovering Stereotypical Gender Patterns 揭示LLM生成文本中性别刻板印象：分析说服性语言的性别差异 large language model
16 ACR: Adaptive Context Refactoring via Context Refactoring Operators for Multi-Turn Dialogue 提出ACR框架，通过自适应上下文重构解决多轮对话中的上下文惯性和状态漂移问题 large language model
17 ReasonAny: Incorporating Reasoning Capability to Any Model via Simple and Effective Model Merging ReasonAny：通过模型融合为任意模型注入推理能力 chain-of-thought
18 Don't Break the Cache: An Evaluation of Prompt Caching for Long-Horizon Agentic Tasks 针对长程Agent任务，评估Prompt缓存以降低成本并提升效率 large language model
19 HAPS: Hierarchical LLM Routing with Joint Architecture and Parameter Search 提出HAPS：联合架构与参数搜索的分层LLM路由框架，提升任务性能。 large language model ✅
20 Left, Right, or Center? Evaluating LLM Framing in News Classification and Generation 评估LLM在新闻分类和生成中的框架效应：揭示普遍存在的意识形态中心塌陷现象 large language model
21 LLMs as Science Journalists: Supporting Early-stage Researchers in Communicating Their Science to the Public 提出一种基于LLM的科学记者框架，辅助早期研究者向公众传播科研成果 large language model
22 Data Augmented Pipeline for Legal Information Extraction and Reasoning 提出一种基于LLM的数据增强流水线，用于提升法律信息抽取与推理性能。 large language model
23 The ICASSP 2026 HumDial Challenge: Benchmarking Human-like Spoken Dialogue Systems in the LLM Era ICASSP 2026 HumDial挑战赛：LLM时代类人对话系统评测基准 large language model
24 CHisAgent: A Multi-Agent Framework for Event Taxonomy Construction in Ancient Chinese Cultural Systems 提出CHisAgent多智能体框架，用于构建古汉语文化系统中的事件分类体系。 large language model
25 FlashMem: Distilling Intrinsic Latent Memory via Computation Reuse FlashMem：通过计算复用提炼LLM的内在隐式记忆，提升长时程自主性。 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🔗⭐
26 CLewR: Curriculum Learning with Restarts for Machine Translation Preference Learning CLewR：一种带重启的课程学习方法，用于机器翻译偏好学习 preference learning curriculum learning large language model ✅
27 Closing the Modality Reasoning Gap for Speech Large Language Models 提出TARS框架，通过强化学习对齐语音大模型中的模态推理差距 reinforcement learning reward design large language model
28 Do LLMs Need Inherent Reasoning Before Reinforcement Learning? A Study in Korean Self-Correction 研究表明：强化学习提升韩语LLM推理能力需先对齐神经元 reinforcement learning large language model
29 EnvScaler: Scaling Tool-Interactive Environments for LLM Agent via Programmatic Synthesis EnvScaler：通过程序化合成扩展LLM Agent工具交互环境 reinforcement learning large language model ✅
30 Chaining the Evidence: Robust Reinforcement Learning for Deep Search Agents with Citation-Aware Rubric Rewards 提出Citation-aware Rubric Rewards，增强LLM深度搜索代理的鲁棒性。 reinforcement learning ✅
31 Pantagruel: Unified Self-Supervised Encoders for French Text and Speech Pantagruel：用于法语文本和语音的统一自监督编码器 representation learning multimodal
32 GIFT: Games as Informal Training for Generalizable LLMs GIFT：利用游戏作为通用LLM的非正式训练环境，提升其泛化能力 reinforcement learning large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页