paper_type.json (366B)
1 { 2 "paper_type": "empirical", 3 "reason": "The paper develops an LLM-based hypothesis generation framework (ACCELMAT) and validates it with quantitative experimental results on the MATDESIGN benchmark, showing improvements over baselines (80% vs 70%, 89% vs 79.67%), with the primary contribution being the experimental findings rather than the benchmark itself." 4 }