ResearchBenchmarksItem-level Evaluation on Item-level QA datasetFollow62.4ROUGE-1LENS14.35226.82639.351.774Dec 28, 2025Evaluation ResultsMethodMethodLinksROUGE-1ROUGE-2ROUGE-LBLEU-4METEORBERTScoreAlignment ScoreLENSModel=Qwen2.5-14B-Inst...Model=Qwen2.5-14B-Instruct2025.1262.44660.338.361.183.273.2TS-TextModel=Qwen2.5-14B-Inst...Model=Qwen2.5-14B-Instruct2025.1217.64.714.21.721.860.466.5TS-ImageModel=Qwen2.5-VL-32BModel=Qwen2.5-VL-32B2025.1216.24.413.61.621.160.268.7