LLM4CodeRE: Generative AI for Code Decompilation Analysis and Reverse Engineering

arXiv:2604.06095v1 Announce Type: cross Abstract: Code decompilation analysis is a fundamental yet challenging task in malware reverse engineering, particularly due to the pervasive use of sophisticated obfuscation techniques. Although recent large language models (LLMs) have shown promise in translating low-level representations into high-level source code, most existing approaches rely on generic code pretraining and lack adaptation to malicious software. We propose LLM4CodeRE, a domain-adapti