SEP 9, 2026 · PREPRINT
Training Trajectories Determine Circuit Removability in Annealable Soft-Prior Transformers
arXiv
Mechanistic study of neural network circuit learning using controlled architectural interventions and ablations; reports reproducible effects but limited to synthetic tasks with small models and no comparison to established baselines or clinical relevance.
Reported
Associative recall accuracy with…0.772 ± 0.020
Associative recall accuracy at ze…0.095 ± 0.009
Associative recall accuracy at ze…0.734 ± 0.028