cs.CL（2025-03-23）

📊 共 15 篇论文 | 🔗 2 篇有代码

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (11 🔗1) 支柱二：RL算法与架构 (RL & Architecture) (3) 支柱一：机器人控制 (Robot Control) (1 🔗1)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (11 篇)

#题目一句话要点标签🔗⭐
1 Unmasking Deceptive Visuals: Benchmarking Multimodal Large Language Models on Misleading Chart Question Answering 提出Misleading ChartQA基准以解决误导性图表问答问题 large language model multimodal
2 MathAgent: Leveraging a Mixture-of-Math-Agent Framework for Real-World Multimodal Mathematical Error Detection 提出MathAgent框架，用于真实场景下多模态数学错误检测 large language model multimodal
3 Mind with Eyes: from Language Reasoning to Multimodal Reasoning 综述多模态推理方法，从语言中心到协同推理，为类人认知能力提供借鉴。 multimodal
4 Investigating Recent Large Language Models for Vietnamese Machine Reading Comprehension 针对越南语机器阅读理解，论文提出使用QLoRA高效微调Llama 3和Gemma模型。 large language model ✅
5 An Empirical Study of the Role of Incompleteness and Ambiguity in Interactions with Large Language Models 研究不完整性和歧义性对大语言模型交互的影响，提出神经符号框架建模多轮问答。 large language model
6 STShield: Single-Token Sentinel for Real-Time Jailbreak Detection in Large Language Models STShield：基于单Token哨兵机制的大语言模型实时越狱攻击检测 large language model
7 "Whose Side Are You On?" Estimating Ideology of Political and News Content Using Large Language Models and Few-shot Demonstration Selection 利用大语言模型和少样本演示选择，估计政治和新闻内容的意识形态倾向 large language model
8 ShED-HD: A Shannon Entropy Distribution Framework for Lightweight Hallucination Detection on Edge Devices 提出ShED-HD框架，在边缘设备上高效检测大语言模型幻觉 large language model
9 LakotaBERT: A Transformer-based Model for Low Resource Lakota Language LakotaBERT：为低资源Lakota语定制的Transformer模型 large language model
10 Instructing the Architecture Search for Spatial-temporal Sequence Forecasting with LLM 提出基于LLM指导的时空序列预测NAS方法，提升搜索效率与效果 large language model
11 WindowKV: Task-Adaptive Group-Wise KV Cache Window Selection for Efficient LLM Inference 提出WindowKV，一种任务自适应的分组KV缓存窗口选择方法，用于高效LLM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🔗⭐
12 Inverse Reinforcement Learning with Dynamic Reward Scaling for LLM Alignment 提出DR-IRL，通过动态奖励调整提升LLM安全对齐效果 reinforcement learning inverse reinforcement learning large language model
13 Understanding the Effects of RLHF on the Quality and Detectability of LLM-Generated Texts 研究表明RLHF虽提升LLM文本质量，但也使其更易被检测且产生冗长重复内容 reinforcement learning RLHF large language model
14 $D^2LoRA$: Data-Driven LoRA Initialization for Low Resource Tasks 提出D²LoRA，一种数据驱动的LoRA初始化方法，提升低资源任务下的微调效率。 DPO direct preference optimization large language model

🔬 支柱一：机器人控制 (Robot Control) (1 篇)

#题目一句话要点标签🔗⭐
15 GeoBenchX: Benchmarking LLMs in Agent Solving Multistep Geospatial Tasks GeoBenchX：评估LLM在多步骤地理空间任务中Agent工具调用能力的基准 manipulation large language model ✅

⬅️ 返回 cs.CL 首页 · 🏠 返回主页