SecRL-Prune: Structured Reinforcement Learning-Based Pruning of CodeLLMs for Preserving Adversarial Code Mutation | ArxivCSExplorer