paper_type.json (395B)
1 { 2 "paper_type": "empirical", 3 "reason": "The paper implements a specific RAG chatbot system and reports quantitative experimental results from testing against 674 adversarial prompts, measuring defense effectiveness (45% recall, 95% benign accuracy), with the primary contribution being the empirical findings about defense mechanisms rather than a novel benchmark or conceptual framework." 4 }