CodeGuard: Improving LLM Guardrails in CS EducationAccepted
Nishat Raihan, Noah Erdachew, Jayoti Devim, Joanna C. S. Santos, Marcos Zampieri
Findings of the 19th Conference of the European Chapter of the ACL (EACL 2026)
A taxonomy, an 8K-prompt dataset, and PromptShield (fine-tuned RoBERTa, 0.93 F1) that cuts harmful code completions by 30–65% while preserving legitimate coding assistance.