ResearchBenchmarksWSI Report Generation on In-house dataset Chinese private (test)Follow0.0924BLEU-1QCAgent0.0198080.0386540.05750.076346Mar 2, 2026Evaluation ResultsMethodMethodLinksBLEU-1BLEU-4ROUGE-LMETEORField RecallBERTScoreQCAgentSetup=Ours, Iterative...Setup=Ours, Iterative QC=true, Iterations (T)=32026.030.09240.01860.2490.23070.48630.237PRISMSetup=Baseline, Iterat...Setup=Baseline, Iterative QC=false2026.030.02260.00830.04840.03330.1908-0.0831