Structured Sparse Memory for Recurrent Reasoning
Recurrent models trained from scratch have recently become competitive on ARC-style reasoning tasks, but the usual framing around small recurrent backbones overlooks two important parts of the system: task-conditioned memory and synthetic augmentation data. We study this regime through CHARM, a compact hybrid ARC model...
Zi-Xuan Zhao, Samuel Wheeler, Neil Getty et al.
· 0 citations