List your company

Knowledge Distillation on The Pile

1,200Raw KL DivergenceStudent
1,139.881,545.691,951.52,357.31Apr 5, 2026

Evaluation results

MethodLinks
StudentdS=1024, Params=405M2026.041,2000.58644.4
StudentdS=768, Params=247M2026.041,3350.65249.4
StudentdS=512, Params=127M2026.041,5010.73355.5
StudentdS=256, Params=45M2026.042,0641.00876.4
StudentdS=128, Params=18M2026.042,7031.32100