Beware What You Autocomplete: Forensic Attribution of Backdoored Code Completions
For security practitioners and developers using LLM-based code completion, CodeTracer provides a practical post-deployment tool to trace backdoor attacks to their source, addressing a critical gap in defense against adaptive attacks.
CodeTracer is a forensic framework that attributes malicious code completions from backdoored LLMs to the specific fine-tuning data responsible, achieving high accuracy and low false identification rates across ten backdoor attacks and sixteen baselines.
Large language models have enabled powerful code completion systems that assist developers by predicting subsequent lines of code. However, these models remain vulnerable to backdoor attacks, where malicious fine-tuning data covertly implants unsafe behaviors. Despite advances in defensive techniques, adaptive and sophisticated backdoor attacks still evade detection and mitigation. We present CodeTracer, a forensic framework that traces malicious code completions back to the backdoor fine-tuning data responsible for them. Operating under realistic post-deployment constraints, CodeTracer relies solely on the fine-tuning corpus and the reported miscompletion event. It extracts a structured behavioral fingerprint from the compromised output, narrows the search to semantically relevant code samples, and employs LLM-based reasoning to attribute unsafe logic to specific backdoor data. Extensive evaluations across three representative vulnerability cases and ten backdoor attacks, along with sixteen competitive baselines, demonstrate that CodeTracer consistently achieves high forensic accuracy, low false identification rates, and strong robustness against adaptive attacks.