Provenance Tracking for LLM-Generated Code: Efficient Attribution at Scale
arXiv·medium signal
Addresses the legal and ethical challenge of LLM code generation reproducing training examples verbatim without attribution. Introduces efficient, scalable provenance tracking that can identify when generated code snippets match existing licensed codebases. Growing regulatory relevance as AI-generated code enters production at scale.