cs.AI（2025-02-03）

📊 共 9 篇论文 | 🔗 3 篇有代码

🎯 兴趣领域导航

#	题目	一句话要点	标签	🔗
1	PSSD: Making Large Language Models Self-denial via Human Psyche Structure	PSSD：通过人类心理结构实现大语言模型的自我否定，提升推理准确性	large language model
2	From Divergence to Consensus: Evaluating the Role of Large Language Models in Facilitating Agreement through Adaptive Strategies	提出基于LLM的自适应协商框架，促进群体决策达成共识	large language model
3	Skewed Memorization in Large Language Models: Quantification and Decomposition	量化并分解大语言模型中的倾斜记忆现象，揭示其与训练数据的关系	large language model
4	DeepRAG: Thinking to Retrieve Step by Step for Large Language Models	DeepRAG：提出一种基于马尔可夫决策过程的检索增强生成框架，提升大语言模型的推理能力。	large language model
5	VideoRAG: Retrieval-Augmented Generation with Extreme Long-Context Videos	提出VideoRAG，用于处理和理解超长上下文视频的检索增强生成框架	large language model	✅
6	Learning to Generate Unit Tests for Automated Debugging	UTGen：学习生成单元测试以辅助LLM自动化调试	large language model
7	Picky LLMs and Unreliable RMs: An Empirical Study on Safety Alignment after Instruction Tuning	揭示指令微调后大语言模型安全性下降问题，并分析奖励模型在安全对齐中的局限性	large language model	✅

#	题目	一句话要点	标签	🔗	⭐
8	TeLL-Drive: Enhancing Autonomous Driving with Teacher LLM-Guided Deep Reinforcement Learning	TeLL-Drive：利用教师LLM引导的深度强化学习增强自动驾驶能力	reinforcement learning deep reinforcement learning DRL
9	Fine-Tuning Discrete Diffusion Models with Policy Gradient Methods	提出SEPO算法，用于策略梯度微调离散扩散模型以解决奖励优化难题	reinforcement learning RLHF	✅