Universal Reasoning Model

Enhance Universal Transformers for complex reasoning through ConvSwiGLU modules integrating depthwise convolution into feed-forward blocks and truncated backpropagation through loops (TBPTL) restricting gradient computation to final iterations. Achieve state-of-the-art on ARC-AGI: 53.8% on ARC-1, 16.0% on ARC-2.

adu2021 Updated 7 repo stars

File contents

adu2021/skillxiv commit f0014704c0

Frequently asked questions

npx get-skill add adu2021/universal-reasoning-model