Breaking Memorization Barriers in LLM Code Fine-Tuning via Information Bottleneck for Improved Generalization

Open in new window