摘要：语境脱离（Out-of-context，OOC）虚假信息将真实图像与具有误导性的标题或说明文字配对，在不篡改图像的情况下构建虚假叙事。因此，其检测问题属于多模态对齐，而非图像取证。尽管 OOC 虚假信息在尼泊尔较为普遍且后果严重，目前尚不存在公开的尼泊尔语基准。我们推出 NepOOC，这是首个公开可用、以尼泊尔语为主的多语言 OOC 基准，包含 1,090 个图文对（545 个原始样本、545 个 OOC 样本），并标注为五种类型：捏造、错误配文、时间不匹配、地理位置不匹配和身份不匹配；标注者间一致性为 kappa = 0.84。我们系统评估了五种多模态架构，以及仅文本和仅图像基线模型。结果显示，在当前数据集规模下，标题文本的语义信息似乎已足以实现强劲性能。仅文本的 mBERT 模型取得 94.65+/-0.20% 的 Macro-F1，与最佳多模态系统 ResNet-50+mBERT（94.65+/-0.20%；McNemar 中位数 p = 1.000；5 个随机种子中 0 个在 alpha = 0.05 下显著）在统计上等效。仅图像模型的表现接近随机水平（33-50%），而训练规模扩展分析表明，扩充数据集是比提升架构复杂度或区域专门化更直接的进步路径。

Abstract:Out-of-context (OOC) misinformation pairs authentic images with misleading captions to construct false narratives without image manipulation, making detection a problem of multimodal alignment rather than image forensics. Despite the prevalence and consequences of OOC misinformation in Nepal, no public benchmark exists for Nepali. We introduce NepOOC, the first publicly available Nepali-dominant multilingual OOC benchmark, comprising 1,090 image-caption pairs (545 pristine, 545 OOC) annotated across five typologies (fabricated, miscaptioned, temporal mismatch, geographic mismatch, identity mismatch) with inter-annotator agreement kappa = 0.84. Systematic evaluation of five multimodal architectures alongside text-only and image-only baselines reveals that caption semantics appear sufficient for strong performance at the current dataset scale. A text-only mBERT model achieves 94.65+/-0.20% Macro-F1, statistically equivalent to the best multimodal system (ResNet-50+mBERT, 94.65+/-0.20%; McNemar median p = 1.000, 0/5 seeds significant at alpha = 0.05). Image-only models perform near chance (33-50%), while training-size scaling suggests that dataset expansion is a more direct path to progress than architectural sophistication or regional specialisation.