cs.CL（2025-03-11）

📊 共 26 篇论文 | 🔗 1 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (21 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱一：机器人控制 (Robot Control) (1) 支柱四：生成式动作 (Generative Motion) (1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (21 篇)

#题目一句话要点标签🔗⭐
1 Cross-Examiner: Evaluating Consistency of Large Language Model-Generated Explanations 提出Cross-Examiner，用于评估大型语言模型生成解释的一致性 large language model
2 Exploring the Word Sense Disambiguation Capabilities of Large Language Models 探索大型语言模型在词义消歧任务中的能力 large language model
3 Exploiting Instruction-Following Retrievers for Malicious Information Retrieval 揭示指令跟随检索器在恶意信息检索中的安全风险 instruction following
4 Position-Aware Depth Decay Decoding ($D^3$): Boosting Large Language Model Inference Efficiency 提出位置感知深度衰减解码以提升大语言模型推理效率 large language model
5 Enhancing Multi-Hop Fact Verification with Structured Knowledge-Augmented Large Language Models 提出LLM-SKAN模型，利用结构化知识增强LLM在多跳事实核查中的性能 large language model
6 Large Language Models for Outpatient Referral: Problem Definition, Benchmarking and Challenges 针对智能门诊转诊，提出基于大语言模型的评估框架与基准测试 large language model
7 CLEV: LLM-Based Evaluation Through Lightweight Efficient Voting for Free-Form Question-Answering CLEV：基于LLM的高效投票评估框架，用于自由形式问答 large language model instruction following
8 LLMs Know What to Drop: Self-Attention Guided KV Cache Eviction for Efficient Long-Context Inference 提出SAGE-KV，利用自注意力指导KV缓存淘汰，提升长文本LLM推理效率。 large language model
9 Interpretable and Robust Dialogue State Tracking via Natural Language Summarization with LLMs 提出基于LLM的自然语言对话状态跟踪(NL-DST)，提升开放域对话的鲁棒性和可解释性。 large language model
10 NSF-SciFy: Mining the NSF Awards Database for Scientific Claims NSF-SciFy：构建大规模科研声明数据集，用于科学发现和评估 large language model
11 DeepReview: Improving LLM-based Paper Review with Human-like Deep Thinking Process DeepReview：通过模拟人类深度思考过程改进基于LLM的论文评审 large language model
12 Transferring Extreme Subword Style Using Ngram Model-Based Logit Scaling 提出基于Ngram模型Logit缩放的极端Subword风格迁移方法，提升大语言模型风格控制能力。 large language model
13 ESPnet-SDS: Unified Toolkit and Demo for Spoken Dialogue Systems ESPnet-SDS：用于语音对话系统的统一工具包与演示平台 foundation model ✅
14 ReviewAgents: Bridging the Gap Between Human and AI-Generated Paper Reviews 提出ReviewAgents框架，利用LLM生成高质量学术论文评审意见，缩小与人类评审的差距。 large language model
15 Fact-checking with Generative AI: A Systematic Cross-Topic Examination of LLMs Capacity to Detect Veracity of Political Information 系统性评估大型语言模型在政治信息核查中的能力与局限性 large language model
16 OpenRAG: Optimizing RAG End-to-End via In-Context Retrieval Learning OpenRAG：通过上下文检索学习端到端优化RAG，提升检索一致性。 large language model
17 Automating Violence Detection and Categorization from Ancient Texts 利用大型语言模型自动检测和分类古代文本中的暴力行为 large language model
18 RigoChat 2: an adapted language model to Spanish using a bounded dataset and reduced hardware RigoChat 2：利用有限数据集和低硬件资源，为西班牙语定制优化语言模型。 large language model
19 OASIS: Order-Augmented Strategy for Improved Code Search 提出OASIS：一种基于排序增强策略的代码搜索方法，提升代码嵌入质量。 large language model
20 Odysseus Navigates the Sirens' Song: Dynamic Focus Decoding for Factual and Diverse Open-Ended Text Generation 提出动态焦点解码（DFD），无需额外数据即可提升开放域文本生成的事实性和多样性。 large language model
21 Learning to Search Effective Example Sequences for In-Context Learning 提出基于Beam Search的示例序列构造器(BESC)，用于优化上下文学习中的示例选择。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (3 篇)

#题目一句话要点标签🔗⭐
22 Backtracking for Safety 提出基于回溯的大语言模型安全对齐方法，解决生成过程中出现的隐蔽性有害内容问题。 reinforcement learning large language model
23 Contrastive Speaker-Aware Learning for Multi-party Dialogue Generation with LLMs 提出Speaker-Attentive LLM，通过对比学习提升多方对话生成质量。 contrastive learning large language model
24 In Prospect and Retrospect: Reflective Memory Management for Long-term Personalized Dialogue Agents 提出反射式记忆管理（RMM）机制，用于提升长期个性化对话Agent的性能。 reinforcement learning large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
25 Dialogue Injection Attack: Jailbreaking LLMs through Context Manipulation 提出对话注入攻击（DIA），利用对话历史破解大型语言模型 manipulation large language model

🔬 支柱四：生成式动作 (Generative Motion) (1 篇)

#题目一句话要点标签🔗⭐
26 Understanding the Quality-Diversity Trade-off in Diffusion Language Models 利用无分类器指导和随机钳位，提升扩散语言模型在序列生成任务中的质量-多样性平衡。 classifier-free guidance

⬅️ 返回 cs.CL 首页 · 🏠 返回主页