About This Repo The code for 263 final project This repo is based on the ActiveIT in the paper Active Instruction Tuning: Improving Cross-Task Generalization by Training on Prompt Sensitive Tasks.