ResearchBenchmarksSummary-level Evaluation on Narrative dataset Summary-level QAFollow59.3ROUGE-1LENS28.20436.27744.3552.423Dec 28, 2025Evaluation ResultsMethodMethodLinksROUGE-1ROUGE-2ROUGE-LBLEU-4METEORBERTScoreCoverageAlignmentLENSModel=Qwen2.5-14B-Inst...Model=Qwen2.5-14B-Instruct2025.1259.326.5412246.777.675.360.1TS-ImageModel=Qwen2.5-VL-32BModel=Qwen2.5-VL-32B2025.1255.621.237.316.645.276.57457.9TS-TextModel=Qwen2.5-14B-Inst...Model=Qwen2.5-14B-Instruct2025.1229.45.715.13.221.863.161.437.2