Preprint
Aug 2026
Beyond Full-Model Rollback: AuroSFT for Adapter-State Multi-Task Fine-Tuning
AuroSFT is introduced, a parameter-efficient framework that recasts the carried state of overfitting-aware multi-task SFT as a compact, mergeable adapter state and obtains higher accuracy on all five backbones.
Yue Han, Zi-Niu Liu
· 0 citations