MultAttnAttrib：长文档问答中的无训练多模态归因
MultAttnAttrib: Training-Free Multimodal Attribution in Long Document Question Answering
AI总结：
提出无训练归因方法MultAttnAttrib，利用模型预填充、选定注意力头和校准阈值定位证据，并构建多模态归因基准MultAttrEval，实验表明其优于多种方法，匹配GPT 5.4且延迟降低至1/7。
摘要：
随着基于事实的问答系统越来越多地部署在AI助手中，准确地将生成的答案归因于证据对于用户信任和模型安全至关重要。虽然单模态归因已被深入探索，但多模态设置仍相对研究不足。因此，我们引入了MultAttnAttrib，一种无需训练的归因生成方法，它利用模型的预填充过程、选定的注意力头和校准阈值来定位文档中的源证据。为了建立该方法的基线结果，我们引入了MultAttrEval，一个互补的基准数据集，其中包含基于多模态源文档的答案组件的细粒度真实归因注释。据我们所知，这是第一个专门为长文档中的多模态归因设计的评估数据集。实验结果表明，MultAttnAttrib始终优于各种归因生成方法，包括几种强大的基于提示的方法，并匹配了最新的前沿模型如GPT 5.4。我们的方法不仅显著提高了单模态和多模态归因类型的归因准确性，而且与在同一基础模型上使用提示相比，归因生成的直接推理延迟降低至七分之一。
英文摘要：
As grounded QA systems are increasingly deployed in AI assistants, accurately attributing generated answers to evidence is critical for user trust and model safety. While unimodal attributions have been explored in depth, the multimodal setting remains relatively under-researched. As a result, we introduce MultAttnAttrib, a training-free attribution-generation method that leverages a model's prefill pass, selected attention heads, and calibrated thresholds to locate source evidence within a document. To establish baseline results for the method, we introduce MultAttrEval, a complementary benchmark dataset annotated with fine-grained, ground-truth attributions for answer components grounded in multimodal source documents. To our knowledge, this is the first evaluation dataset designed specifically for multimodal attribution in long-form documents. Experimental results show that MultAttnAttrib consistently outperforms a variety of attribution-generation methods, including several strong prompting-based approaches and matches the latest frontier models such as GPT 5.4. Our method not only substantially improves attribution accuracy for both unimodal and multimodal attribution types, but also produces attributions at up to one-seventh of the direct inference latency compared to prompting on the same base model.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
