cs.CL（2024-07-23）

📊 共 20 篇论文

🎯 兴趣领域导航

支柱九：具身大模型 (Embodied Foundation Models) (15) 支柱二：RL算法与架构 (RL & Architecture) (5)

🔬 支柱九：具身大模型 (Embodied Foundation Models) (15 篇)

#题目一句话要点标签🔗⭐
1 Do LLMs Know When to NOT Answer? Investigating Abstention Abilities of Large Language Models 提出Abstain-QA数据集与黑盒评估方法，研究大语言模型的回避回答能力。 large language model chain-of-thought
2 AMONGAGENTS: Evaluating Large Language Models in the Interactive Text-Based Social Deduction Game AMONGAGENTS：利用大型语言模型在互动文本社交推理游戏中评估智能体行为 large language model
3 Machine Translation Hallucination Detection for Low and High Resource Languages using Large Language Models 利用大型语言模型检测机器翻译中的幻觉，提升低资源和高资源语言翻译质量 large language model
4 Structure-aware Domain Knowledge Injection for Large Language Models 提出StructTuning，利用结构化领域知识高效微调大语言模型，仅需5%数据达到传统知识注入效果。 large language model
5 An Active Inference Strategy for Prompting Reliable Responses from Large Language Models in Medical Practice 提出基于主动推理的LLM提示策略，提升医疗场景下LLM响应的可靠性 large language model
6 Robust Privacy Amidst Innovation with Large Language Models Through a Critical Assessment of the Risks 通过风险评估，利用大语言模型在创新中实现稳健的隐私保护 large language model
7 Finetuning Generative Large Language Models with Discrimination Instructions for Knowledge Graph Completion DIFT：通过判别指令微调生成式大语言模型，用于知识图谱补全。 large language model
8 Generation Constraint Scaling Can Mitigate Hallucination 提出生成约束缩放方法，无需训练即可缓解记忆增强型LLM中的幻觉问题 large language model
9 Retrieval Augmented Generation or Long-Context LLMs? A Comprehensive Study and Hybrid Approach 提出Self-Route方法，根据模型自反思动态选择RAG或长文本LLM，降低计算成本并保持性能。 large language model
10 Lawma: The Power of Specialization for Legal Annotation Lawma：利用专业化提升法律文本标注性能 large language model
11 TookaBERT: A Step Forward for Persian NLU TookaBERT：面向波斯语NLU的BERT模型，显著提升性能 foundation model
12 LawLuo: A Multi-Agent Collaborative Framework for Multi-Round Chinese Legal Consultation LawLuo：多智能体协作框架，用于多轮中文法律咨询 large language model
13 PreAlign: Boosting Cross-Lingual Transfer by Early Establishment of Multilingual Alignment PreAlign：通过提前建立多语言对齐来提升跨语言迁移性能 large language model
14 Graph-Structured Speculative Decoding 提出图结构推测解码（GSD）加速LLM推理，显著提升token接受率和推理速度。 large language model
15 How to Leverage Personal Textual Knowledge for Personalized Conversational Information Retrieval 利用个人文本知识，通过大语言模型进行个性化对话式信息检索。 large language model

🔬 支柱二：RL算法与架构 (RL & Architecture) (5 篇)

#题目一句话要点标签🔗⭐
16 A Comprehensive Survey of LLM Alignment Techniques: RLHF, RLAIF, PPO, DPO and More 全面综述LLM对齐技术：RLHF、RLAIF、PPO、DPO等 PPO RLHF DPO
17 Can Large Language Models Automatically Jailbreak GPT-4V? 提出AutoJailbreak，利用LLM自动破解GPT-4V的安全防护，攻击成功率超95.3%。 RLHF large language model multimodal
18 DDK: Distilling Domain Knowledge for Efficient Large Language Models DDK：通过领域知识蒸馏提升高效大语言模型性能 distillation large language model
19 TLCR: Token-Level Continuous Reward for Fine-grained Reinforcement Learning from Human Feedback 提出TLCR以解决人类反馈强化学习中的奖励不匹配问题 reinforcement learning RLHF
20 Course-Correction: Safety Alignment Using Synthetic Preferences 提出基于合成偏好的课程纠正方法，提升大语言模型安全性。 preference learning large language model

⬅️ 返回 cs.CL 首页 · 🏠 返回主页