Nishat Raihan, Noah Erdachewm, Jayoti Devim, Joanna C. S. Santos, Marcos Zampieri
Findings of the 19th Conference of the European Chapter of the ACL (EACL 2026)
AI Safety
A taxonomy, an 8K-prompt dataset, and PromptShield (0.93 F1) that cut harmful code completions by 30–65% while preserving legitimate coding assistance.