Safer Content or Firmer Refusals? A Hybrid Perturbation Defense for Alignment under Harmful Fine-tuning

Muhammad Zeeshan Akram, Mufid Kamel Marican, Anvesh Reddy Yenugu +2

Fine-tuning-as-a-service lets users adapt a safety-aligned language model to their own data, but it also creates a harmful fine-tuning attack surface: a small amount of harmful dat…