作者位次 #7/9ACL 2026Poster· Natural Language Processing
D eep G uard: Secure Code Generation via Multi-Layer Semantic Aggregation
Li Huang · Zhongxin Liu · Yifan Wu · Tao Yin · Dong li · Jichao Bi · Nankun Mu · Hongyu Zhang · Meng Yan
摘要
Large Language Models (LLMs) for code generation can replicate insecure patterns from their training data. To mitigate this, a common strategy for security hardening is to fine-tune models using supervision derived from the final transformer layer. However, this design may suffer from a final-lay...