LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging

Open in new window