Delphi Midtraining Interactive Scaling

Interactive validation-trajectory prediction for math validation loss. Prefix-fit variants include MAE and Huber; final model comparison uses endpoint MAE.

learning rates

Within-Run Prediction

Lines are observed trajectories. Diamonds are observed finals; x markers are predicted finals. Parametric curve methods draw fitted continuations from the selected prefix to the final step.

Best Parametric Form By Scale And Prefix

Each cell shows the parametric curve variant with lowest final-loss MAE at that scale and prefix.

Configs Meeting Target MAE For Every Completed Run

A config qualifies only if every completed run in the group has absolute final-loss error at or below the selected target. Tables are sorted by smallest prefix, then worst-case error.

Overall

Per Scale