4. 迷失在末尾:多模态检索增强问答中的首因偏差
研究发现,recall@k 并非已部署 KB-VQA 的正确评价指标,且弥合差距需要 reader 侧介入;本文首次对多模态 KB-VQA 中 reader 侧位置依赖性进行了受控探查,设计了一种 gold-position 协议——在问题提示中仅改变 gold passage 所在的槽位。The findings indicate that recall@k is the wrong metric for deployed KB-VQA and that closing the gap requires reader-side intervention; the first controlled probe of reader-side position dependence in multimodal KB-VQA is designed, a gold-position protocol in which only the gold passage's prompt slot varies within question.