Transformer Architecture

2 papers touch this idea.

Papers

One AI brain doing two jobs at once? Split it, and it learns better.

One Transformer Layer Can Match Full RL Training

Related concepts

Ideas that show up alongside Transformer Architecture.

Agentic Decision-Making Code Generation Data And Compute Efficiency Gradient Analysis Language Modeling Large Language Model Post-Training Layer-Wise Analysis Mathematical Reasoning Model Pretraining Parameter-Efficient Training Reinforcement Learning State-Prediction Separation