cs.CL（2025-12-17）

📊 共 17 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) (14) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (14 篇)

#	题目	一句话要点	标签	🔗	⭐
1	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
2	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服务器环境下的安全性	large language model
3	Dual-Density Inference for Efficient Language Model Reasoning	提出Denser双密度推理框架，提升LLM在复杂推理问答任务中的效率。	large language model chain-of-thought
4	Activation Oracles: Training and Evaluating LLMs as General-Purpose Activation Explainers	提出Activation Oracles，通过多样化训练提升LLM激活解释的通用能力。	large language model
5	How Much is Too Much? Exploring LoRA Rank Trade-offs for Retaining Knowledge and Domain Robustness	探索LoRA秩对知识保留和领域泛化能力的权衡，优化下游问答任务	large language model
6	The Meta-Prompting Protocol: Orchestrating LLMs via Adversarial Feedback Loops	提出Meta-Prompting协议，通过对抗反馈循环实现LLM的可靠编排与自优化。	large language model
7	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
8	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
9	Yes-MT's Submission to the Low-Resource Indic Language Translation Shared Task in WMT 2024	Yes-MT团队探索多种模型和微调策略，解决WMT 2024低资源印度语言翻译难题。	large language model
10	RFKG-CoT: Relation-Driven Adaptive Hop-count Selection and Few-Shot Path Guidance for Knowledge-Aware QA	RFKG-CoT：关系驱动的自适应跳数选择与少样本路径引导，提升知识图谱问答效果	large language model
11	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的KV缓存检索方法	large language model
12	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
13	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
14	The Moralization Corpus: Frame-Based Annotation and Analysis of Moralizing Speech Acts across Diverse Text Genres	构建道德化语料库，用于分析跨文本类型的道德化言语行为	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	🔗	⭐
15	Well Begun, Half Done: Reinforcement Learning with Prefix Optimization for LLM Reasoning	提出PPPO方法，通过优化LLM推理前缀token策略，提升强化学习推理能力。	reinforcement learning large language model
16	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，揭示其在特定类型信息上的记忆短板。	Mamba SSM state space model
17	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性能	reinforcement learning large language model	✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页