Deletion tuning: how far can a model improve by removing weights?

AI · Language model adaptation and interpretability

Loading…