Understanding On-Policy Distillation: A Mechanistic Interpretability Perspective via Sparse Crosscoders

Open paper
Benchmark Model Rank Results
mathematical-reasoning-on-aime24OPD student (JustRL)#85Acc: 46.9
mathematical-reasoning-on-aime-2025OPD student (JustRL)#105Accuracy: 36.0

Rank counts only results with a code link.