paper_type.json (303B)
1 { 2 "paper_type": "empirical", 3 "reason": "Runs controlled experiments comparing zero-shot vs few-shot LLM repair, evaluates auxiliary information impact, and reports quantitative results (repair rates, percentage improvements/degradations) across multiple models and the proposed SRepair framework." 4 }