AINeutralarXiv – CS AI · 15h ago6/10
🧠
Shadow Unlearning: A Neuro-Semantic Approach to Fidelity-Preserving Faceless Forgetting in LLMs
Researchers introduce Shadow Unlearning, a privacy-preserving machine unlearning method that removes training data influence from LLMs without exposing sensitive information to attacks. The Neuro-Semantic Projector Unlearning (NSPU) framework achieves this while maintaining model performance and is 10x more computationally efficient than existing approaches.