paper_type.json (323B)
1 { 2 "paper_type": "empirical", 3 "reason": "Runs experiments on 900 agentic execution traces across three LLMs, reports quantitative performance metrics (e.g., 2/30 accuracy), and contributes experimental findings about failure archetypes rather than introducing a novel benchmark or dataset as the primary contribution." 4 }