Tag: Reasoning

All the articles with the tag "Reasoning".

• Activation-Guided Consensus Merging for Large Language Models

  Published: 22 May, 2025 |  at  11:19 AM

  

  90.71 🤔

  本文提出Activation-Guided Consensus Merging (ACM)，通过基于激活值互信息（MI）的层级权重系数调整，实现大型语言模型在Long-to-Short推理任务中的高效合并，显著减少输出冗余并提升推理精度，尤其在小规模模型上效果明显。

• Beyond Single-Task: Robust Multi-Task Length Generalization for LLMs

  Published: 23 May, 2025 |  at  11:14 AM

  

  90.65 🤔

  本文提出Meta-RFFT框架，通过多任务规则跟随预训练和少量下游适应，显著提升了大型语言模型在未见任务上的长度泛化能力，32B模型在长度30的加法任务上达到98%准确率，超越现有长链推理模型。

• REARANK: Reasoning Re-ranking Agent via Reinforcement Learning

  Published: 30 May, 2025 |  at  11:19 AM

  

  90.58 🤔

  本文提出REARANK，一种基于强化学习的列表式重排序代理，通过显式推理和数据增强，仅用179个标注查询即在多个信息检索基准上显著超越基线并媲美甚至超越GPT-4，尤其在推理密集型任务中表现突出。

• LiteWebAgent: The Open-Source Suite for VLM-Based Web-Agent Applications

  Published: 14 May, 2025 |  at  11:12 AM

  

  90.54 🤔

  LiteWebAgent is an open-source suite for VLM-based web agents that bridges the gap in production-ready solutions by offering an extensible framework with decoupled action generation and grounding, advanced planning, memory, tree search, and practical deployments via Vercel and Chrome extension.

• Thinking Out Loud: Do Reasoning Models Know When They're Right?

  Published: 25 May, 2025 |  at  11:51 AM

  

  90.51 🤔

  本文通过对比指令微调、监督微调和强化学习训练的大型推理模型，发现推理导向训练显著提升了推理任务中的准确性和校准能力，但在事实性任务中可能削弱小规模模型对知识边界的感知。