Label Smoothing Improves Gradient Ascent in LLM Unlearning

Open in new window