paper_type.json (297B)
1 { 2 "paper_type": "benchmark-creation", 3 "reason": "Introduces UniCode, a generative evaluation framework that augments programming problems along five structured axes, with empirical testing primarily serving to validate and demonstrate the benchmark's utility in revealing model limitations." 4 }