cs.CL（2025-12-17）

📊 共 19 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) (16) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (16 篇)

#题目一句话要点标签🔗⭐
1 Evaluating Large Language Models on Multimodal Chemistry Olympiad Exams 评估大型语言模型在多模态化学奥林匹克竞赛题上的表现 large language model multimodal visual grounding
2 SGM: Safety Glasses for Multimodal Large Language Models via Neuron-Level Detoxification SGM：通过神经元级解毒为多模态大语言模型提供安全保障 large language model multimodal
3 MCP-SafetyBench: A Benchmark for Safety Evaluation of Large Language Models with Real-World MCP Servers 提出MCP-SafetyBench，用于评估大语言模型在真实MCP服务器环境中的安全性 large language model
4 Dual-Density Inference for Efficient Language Model Reasoning 提出Denser双密度推理框架，提升LLM复杂推理任务的计算效率。 large language model chain-of-thought
5 BRAID: Bounded Reasoning for Autonomous Inference and Decisions BRAID：一种用于自主推理和决策的有界推理框架，提升LLM的效率和准确性 large language model
6 Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers 提出Activation Oracles，通过训练LLM解释激活值，实现通用激活解释器。 large language model
7 How Much is Too Much? Exploring LoRA Rank Trade-offs for Retaining Knowledge and Domain Robustness 研究LoRA秩对知识保留和领域泛化能力的权衡，为下游任务提供参数高效微调策略。 large language model
8 Evaluating Metrics for Safety with LLM-as-Judges 提出基于LLM-as-Judges的加权指标评估方法，提升LLM在安全关键任务中的可靠性。 large language model
9 CTkvr: KV Cache Retrieval for Long-Context LLMs via Centroid then Token Indexing 提出CTKVR：一种基于质心和Token索引的长文本LLM KV缓存检索方法 large language model
10 Toward expert-level motivational interviewing for health behavior improvement with LLMs 利用大型语言模型实现专家级动机访谈，促进健康行为改善 large language model
11 Evaluating LLMs for Zeolite Synthesis Event Extraction (ZSEE): A Systematic Analysis of Prompting Strategies 系统评估LLM在沸石合成事件抽取（ZSEE）中的提示策略有效性 large language model
12 Towards Proactive Personalization through Profile Customization for Individual Users in Dialogues 提出PersonalAgent，通过用户画像定制实现对话系统中的主动个性化 large language model
13 The Moralization Corpus: Frame-Based Annotation and Analysis of Moralizing Speech Acts across Diverse Text Genres 提出道德化语料库以分析多样文本中的道德化言论 large language model
14 Yes-MT's Submission to the Low-Resource Indic Language Translation Shared Task in WMT 2024 Yes-MT团队探索多种方法，解决WMT 2024低资源印度语言翻译难题。 large language model
15 RFKG-CoT: Relation-Driven Adaptive Hop-count Selection and Few-Shot Path Guidance for Knowledge-Aware QA 提出RFKG-CoT以解决知识密集型问答中的幻觉问题 large language model
16 The Meta-Prompting Protocol: Orchestrating LLMs via Adversarial Feedback Loops 提出Meta-Prompting协议，通过对抗反馈循环优化LLM，提升可靠性。 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🔗⭐
17 Characterizing Mamba's Selective Memory using Auto-Encoders 利用自编码器剖析Mamba选择性记忆的遗忘特性，揭示其对特定类型信息的遗忘偏好。 Mamba SSM state space model
18 Well Begun, Half Done: Reinforcement Learning with Prefix Optimization for LLM Reasoning 提出PPPO：通过前缀优化强化学习提升LLM的推理能力 reinforcement learning large language model
19 Beyond Majority Voting: Towards Fine-grained and More Reliable Reward Signal for Test-Time Reinforcement Learning 提出SCOPE框架，通过细粒度置信度加权伪标签提升测试时强化学习效果 reinforcement learning large language model ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页