cs.CL（2026-05-18）

📊 共 23 篇论文 | 🔗 4 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (20 🔗3) 支柱二：RL算法与架构 (RL & Architecture) (3 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (20 篇)

#题目一句话要点标签🔗⭐
1 BacktestBench: Benchmarking Large Language Models for Automated Quantitative Strategy Backtesting BacktestBench：用于自动化量化策略回测的大语言模型基准测试平台 large language model
2 Prompt Compression in Diffusion Large Language Models: Evaluating LLMLingua-2 on LLaDA 评估LLMLingua-2在扩散大语言模型LLaDA上的Prompt压缩性能，揭示其与自回归模型的差异 large language model
3 How Good LLMs Are at Answering Bangla Medical Visual Questions? Dataset and Benchmarking 提出BanglaMedVQA数据集，评估LLM在孟加拉语医学视觉问答中的能力 large language model foundation model
4 Code as Agent Harness 提出“代码即Agent Harness”统一视角，研究代码在智能体系统中的核心作用。 large language model multimodal
5 DashAttention: Differentiable and Adaptive Sparse Hierarchical Attention 提出DashAttention，一种可微自适应稀疏分层注意力机制，提升长文本建模效率。 large language model
6 Predictable Confabulations: Factual Recall by LLMs Scales with Model Size and Topic Frequency 揭示LLM事实性知识回忆能力与模型规模及主题频率的scaling law关系 large language model
7 EvoMemBench: Benchmarking Agent Memory from a Self-Evolving Perspective EvoMemBench：从自进化角度评估LLM Agent记忆能力，填补现有评测体系空白 large language model ✅
8 From Volume to Value: Preference-Aligned Memory Construction for On-Device RAG 提出EPIC，通过偏好对齐的内存构建，实现高效的端侧RAG large language model
9 Machine Unlearning for Masked Diffusion Language Models 提出Masked Diffusion Unlearning (MDU)，用于擦除Masked Diffusion语言模型中的特定知识。 large language model ✅
10 Multilingual jailbreaking of LLMs using low-resource languages 利用低资源语言进行多语言LLM越狱攻击研究 large language model
11 Context Memorization for Efficient Long Context Generation 提出Attention-State Memory，解决长文本生成中前缀信息衰减和注意力计算效率问题。 large language model
12 Predictive Prefetching for Retrieval-Augmented Generation 提出预测式预取框架，解决检索增强生成中同步检索带来的延迟问题。 large language model
13 Multi-agent AI systems outperform human teams in creativity 多智能体LLM系统在创造力方面超越人类团队，解决创新难题。 large language model
14 MA$^{2}$P: A Meta-Cognitive Autonomous Intelligent Agents Framework for Complex Persuasion 提出MA²P框架，用于复杂说服场景下提升自主智能体的说服成功率。 large language model
15 Monitoring the Internal Monologue: Probe Trajectories Reveal Reasoning Dynamics 提出探针轨迹以监测大型推理模型的推理动态 chain-of-thought
16 Implicit Hierarchical GRPO: Decoupling Tool Invocation from Execution for Tool-Integrated Mathematical Reasoning 提出隐式分层GRPO算法，解耦工具调用与执行，提升工具集成数学推理能力 large language model ✅
17 Presupposition and Reasoning in Conditionals: A Theory-Based Study of Humans and LLMs 研究条件句中预设与推理，对比人类与大语言模型在语用学上的差异 large language model
18 A Data-Efficient Path to Multilingual LLMs: Language Expansion via Post-training PARAM$Δ$ Integration into Upcycled MoE 提出PARAMΔ集成方法以高效扩展多语言LLM large language model
19 PPAI: Enabling Personalized LLM Agent Interoperability for Collaborative Edge Intelligence PPAI：实现个性化LLM Agent互操作，赋能协同边缘智能 large language model
20 Internalizing Tool Knowledge in Small Language Models via QLoRA Fine-Tuning 通过QLoRA微调，将工具知识内化于小型语言模型中，实现无描述推理。 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🔗⭐
21 Sometin Beta Pass Notin (SBPN): Improving Multilingual ASR for Nigerian Languages via Knowledge Distillation SBPN：通过知识蒸馏提升尼日利亚语多语种自动语音识别性能 distillation foundation model
22 Knowledge-to-Verification: Exploring RLVR for LLMs in Knowledge-Intensive Domains 提出K2V框架，通过可验证数据自动合成增强LLM在知识密集型领域的推理能力。 reinforcement learning large language model ✅
23 AutoVecCoder: Teaching LLMs to Generate Explicitly Vectorized Code AutoVecCoder：教导LLM生成显式向量化代码，提升高性能计算效率 reinforcement learning large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页