Adaptive Defense against Harmful Fine-Tuning for Large Language Models via Bayesian Data Scheduler

AAMLRe-assign community