LLM Unlearning using Gradient Ratio-Based Influence Estimation and Noise Injection

Open in new window