Posts

All the articles I've posted.

• SAGE: A Framework of Precise Retrieval for RAG

  Published: 4 May, 2025 |  at  04:27 PM

  

  64.13 🤔

  本文提出SAGE框架，通过语义分割、基于梯度的块选择和LLM自反馈机制，提高RAG系统的检索精度和问答性能，同时显著降低成本。

• SmallPlan: Leverage Small Language Models for Sequential Path Planning with Simulation-Powered, LLM-Guided Distillation

  Published: 5 May, 2025 |  at  11:15 PM

  

  64.11 🤔

  本文提出了 SmallPlan 框架，通过结合 LLM 指导的蒸馏、模拟环境反馈的 SFT 和 RL，训练轻量级的小型语言模型 (SLM) 进行高效的机器人高层路径规划，使其在资源受限的边缘设备上实现接近大型模型 (LLM) 的性能。

• DyMU: Dynamic Merging and Virtual Unmerging for Efficient VLMs

  Published: 4 May, 2025 |  at  04:29 PM

  

  63.91 🤔

  本文提出DYMU框架，通过动态令牌合并和虚拟取消合并的训练-free方法，显著提高了VLMs的计算效率，同时在多个基准上保持了与完整模型相似的性能。

• The Promise and Limits of LLMs in Constructing Proofs and Hints for Logic Problems in Intelligent Tutoring Systems

  Published: 13 May, 2025 |  at  11:09 AM

  

  63.17 🤔

  This paper evaluates LLMs in intelligent tutoring systems for propositional logic, demonstrating DeepSeek-V3's promising accuracy in proof construction (up to 86.7%) and hint generation (75%), but reveals significant pedagogical limitations in justification and subgoaling, necessitating hybrid approaches for educational integration.

• MAC-Tuning: LLM Multi-Compositional Problem Reasoning with Enhanced Knowledge Boundary Awareness

  Published: 4 May, 2025 |  at  04:32 PM

  

  63.08 🤔

  本文提出MAC-Tuning方法，通过分步微调分离答案预测和置信度估计，提升LLMs在多问题设置下的知识边界意识，显著减少幻觉并改善性能。