Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Grokking Onset and Speedup on 3-digit Addition (NanoGPT Implementation)

9,800Grokking Onset

Norm Penalty

9,29212,72116,15019,579Jun 30, 2026
Updated 25d ago

Evaluation Results

MethodLinks
2026.06
9,8002.3
2026.06
22,500-