paper_type.json (245B)
1 { 2 "paper_type": "empirical", 3 "reason": "The paper experimentally evaluates Astrogator across 6 LLMs on 1260 generated programs, reporting quantitative metrics for verification accuracy and LLM code correctness as its primary contribution." 4 }