README.md

metadata

tags:
  - Distillation

the model(student) after finetuning outperforms 410m-deduped model on wsc accuracy.

lm-eval : aloobun/distill-p-test

lm-eval : EleutherAI/pythia-410m-deduped