Salesforce AI Research 发布了 E-VQA 任务和 ST-Evidence 基准，解决了视频问答无法验证视觉证据的问题。如果你想理解视频模型真正“看到”了什么，这篇论文值得一读。
当前视频大语言模型在问答中表现优异但缺乏可验证的视觉依据。研究人员提出E-VQA（Evidence-Backed Video Question Answering）新任务，要求模型同时输出语义答案和精确的时空证据：时间片段与密集跟踪物体分割掩码。为此构建了首个经人工验证的像素级定位基准ST-Evidence，包含判别与生成两种任务。对现有模型的评估显示，问答准确率与真实视觉感知之间存在显著脱节，仅靠扩展规模无法解决。团队开发了自动化生成流程，创建16万规模的ST-Evidence-Instruct数据集，微调后的7B模型在UniPixel基线上获得+27.2 t-mean和+13.8 J&F的提升。
Evidence-Backed Video Question Answering
Current Video Large Language Models (Video LLMs) excel in question answering (QA) but largely operate as black boxes, providing textual answers without verifiable visual grounding. Existing explainability efforts rely on textual rationales or sparse bounding boxes, which struggle to capture complex video dynamics such as occlusions and non-rigid deformations. We propose Evidence-Backed Video Question Answering (E-VQA), a novel task requiring models to jointly output a semantic answer and precise spatio-temporal evidence: temporal segments and dense, tracked object segmentation masklets. To support this, we introduce ST-Evidence, the first human-verified benchmark for both discriminative and generative pixel-level grounding. Evaluations of state-of-the-art models reveal a critical decoupling between QA accuracy and true visual perception that scaling alone fails to bridge. To address this, we develop scalable, automated generation pipelines to create ST-Evidence-Instruct, a 160k-scale dataset bridging high-level reasoning with fine-grained grounding. Fine-tuning grounded Video LLMs on this data yields substantial gains over the corresponding size-matched UniPixel baselines (e.g., +27.2 t-mean and +13.8 J&F on a 7B model), establishing a robust baseline for explainable, evidence-backed video understanding. Code and data are available at https://github.com/SalesforceAIResearch/EVQA.