ResearchBenchmarksRubric Generation on Research RubricsRubric Generation on Research RubricsFollow13.6BLEURubricRAG8.7129.98111.2512.519Mar 21, 2026Evaluation resultsMethodMethodLinksBLEUROUGE-1ROUGE-2ROUGE-LLLM Judge ScoreRubricRAGThinking=thinkThinking=think2026.0313.633.713.528.352.3Few-ShotMode=Few-ShotMode=Few-Shot2026.0313.333.512.927.550.1Zero-ShotMode=Zero-ShotMode=Zero-Shot2026.039.222.47.11851.4RubricRAGThinking=nothinkThinking=nothink2026.038.933.913.328.657.6