cs.CL（2025-09-09）

📊 共 24 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) (4 🔗1)

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

#题目一句话要点标签🔗⭐
1 Dual Knowledge-Enhanced Two-Stage Reasoner for Multimodal Dialog Systems 提出DK2R模型，利用双重知识增强多模态对话系统中的文本回复生成。 large language model multimodal
2 Bias after Prompting: Persistent Discrimination in Large Language Models 揭示Prompting后的大语言模型偏见：持续存在的歧视现象 large language model
3 GENUINE: Graph Enhanced Multi-level Uncertainty Estimation for Large Language Models GENUINE：图增强多层次不确定性估计，提升大语言模型可靠性 large language model ✅
4 Automated Item Neutralization for Non-Cognitive Scales: A Large Language Model Approach to Reducing Social-Desirability Bias 利用大型语言模型减少个性评估中的社会期望偏差 large language model
5 Are Humans as Brittle as Large Language Models? 对比人类与大语言模型，揭示提示词修改对文本分类任务的影响 large language model
6 M-BRe: Discovering Training Samples for Relation Extraction from Unlabeled Texts with Large Language Models 提出M-BRe框架，利用大语言模型从无标注文本中高效挖掘关系抽取训练样本 large language model
7 NOWJ@COLIEE 2025: A Multi-stage Framework Integrating Embedding Models and Large Language Models for Legal Retrieval and Entailment NOWJ团队提出多阶段框架，融合嵌入模型与大语言模型，用于法律检索和蕴含任务。 large language model
8 LongEmotion: Measuring Emotional Intelligence of Large Language Models in Long-Context Interaction LongEmotion：提出用于评估大语言模型在长文本交互中情感智能的基准 large language model ✅
9 Are LLMs Enough for Hyperpartisan, Fake, Polarized and Harmful Content Detection? Evaluating In-Context Learning vs. Fine-Tuning 对比In-Context Learning与微调，评估大语言模型在检测有害内容方面的能力。 large language model chain-of-thought
10 VeriOS: Query-Driven Proactive Human-Agent-GUI Interaction for Trustworthy OS Agents 提出VeriOS，通过查询驱动的人机交互提升OS Agent在不可信环境下的可靠性 large language model multimodal ✅
11 Dynamic Prompt Fusion for Multi-Task and Cross-Domain Adaptation in LLMs 提出动态Prompt融合框架，提升LLM在多任务和跨领域场景下的泛化能力 large language model
12 Evolution and compression in LLMs: On the emergence of human-aligned categorization 研究表明，大型语言模型可以通过迭代学习进化出与人类对齐的语义分类系统。 large language model
13 No for Some, Yes for Others: Persona Prompts and Other Sources of False Refusal in Language Models 研究表明：人格化提示可能导致LLM产生虚假拒绝，但影响程度或被高估 large language model
14 SimpleQA Verified: A Reliable Factuality Benchmark to Measure Parametric Knowledge 提出SimpleQA Verified，用于可靠评估LLM参数知识的事实性，并缓解幻觉问题。 large language model
15 Biased Tales: Cultural and Topic Bias in Generating Children's Stories Biased Tales：揭示并分析LLM生成儿童故事中的文化和主题偏见 large language model
16 From Detection to Mitigation: Addressing Gender Bias in Chinese Texts via Efficient Tuning and Voting-Based Rebalancing 提出基于LoRA微调和投票机制的中文性别偏见检测与缓解方法 large language model
17 ALLabel: Three-stage Active Learning for LLM-based Entity Recognition using Demonstration Retrieval 提出ALLabel，一种基于演示检索的三阶段主动学习框架，用于提升LLM在实体识别中的性能。 large language model
18 Talking with Oompa Loompas: A novel framework for evaluating linguistic acquisition of LLM agents 提出Tinkatongue框架，评估LLM智能体在交互中学习新语言的能力 large language model
19 PersonaFuse: A Personality Activation-Driven Framework for Enhancing Human-LLM Interactions PersonaFuse：基于人格激活的框架，增强人与LLM的交互 large language model
20 Does This Look Familiar to You? Knowledge Analysis via Model Internal Representations 提出KAMIR方法，通过模型内部表征分析进行高效训练数据选择，提升模型泛化能力。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (4 篇)

#题目一句话要点标签🔗⭐
21 SciGPT: A Large Language Model for Scientific Literature Understanding and Knowledge Discovery SciGPT：面向科学文献理解和知识发现的大语言模型 distillation large language model foundation model
22 MVPBench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values MVPBench：构建基准与微调框架，对齐大语言模型与多元人类价值观 DPO direct preference optimization large language model
23 Parallel-R1: Towards Parallel Thinking via Reinforcement Learning 提出Parallel-R1框架，通过强化学习赋能LLM并行思维能力，提升复杂推理任务性能。 reinforcement learning large language model ✅
24 MERLIN: Multi-Stage Curriculum Alignment for Multilingual Encoder-LLM Integration in Cross-Lingual Reasoning MERLIN：多阶段课程对齐框架，提升跨语言推理中多语言编码器-LLM集成效果 curriculum learning large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页