Tag: Emergent Abilities

All the articles with the tag "Emergent Abilities".

• SimpleRL-Zoo: Investigating and Taming Zero Reinforcement Learning for Open Base Models in the Wild

  Published: 8 May, 2025 |  at  06:17 PM

  

  86.55 🤔

  This paper investigates zero RL training on diverse open base models, achieving significant accuracy and response length improvements while identifying key factors like reward design and data difficulty that influence the emergence of reasoning behaviors.

• ICLR: In-Context Learning of Representations

  Published: 7 May, 2025 |  at  08:41 AM

  

  84.18 🤔

  本文通过上下文图追踪任务揭示了大型语言模型能随上下文规模增加而突现地重组概念表示以适应新语义，并提出能量最小化假设解释这一过程。

• Toward Understanding In-context vs. In-weight Learning

  Published: 7 May, 2025 |  at  12:16 AM

  

  82.20 🤔

  本文通过一个简化的理论模型和多场景实验，揭示了数据分布特性如何驱动上下文学习（ICL）和权重学习（IWL）的出现与竞争，并解释了ICL在训练过程中可能短暂的原因。