Large language model safety evaluation remains heavily English-centered, leaving low-resource languages under-measured even when models are deployed globally. We evaluate four open-weight instruction-tuned models on SomaliBench v0, a native-author-verified benchmark of 100 harmful-intent prompts paired across English and Somali. Each of Llama-3.1-8B-Instruct, Gemma-2-9B-Instruct, Qwen-2.5-7B-Instruct, and Aya-23-8B is run locally with temperature 0 and the same English "helpful, harmless, and honest" (HHH) system prompt. A pinned Claude Sonnet snapshot (claude-sonnet-4-5-20250929) classifies each response as refused, complied, or unclear; the native author spot-checks a stratified 80-row sample. We find large English-to-Somali refusal gaps for all four models: Llama-3.1-8B (0.90; 95% bootstrap CI [0.85, 0.96]), Aya-23-8B (0.75 [0.67, 0.83]), Qwen-2.5-7B (0.69 [0.59, 0.78]), and Gemma-2-9B (0.38 [0.27, 0.49]). For three models, the dominant Somali non-refusal mode is not fluent harmful compliance but unclear output: empty, wrong-language, or incoherent generations. The native verification spot-check achieves 100% agreement with the judge (Cohen's kappa = 1.00) on the 80 sampled rows. We report aggregate refusal rates, category gaps, and reliability statistics only; raw model generations are retained locally and are not released.

翻译：大型语言模型的安全性评估仍高度以英语为中心，导致低资源语言在模型全球部署中缺乏充分测量。我们对四个开源权重指令微调模型在SomaliBench v0（由母语者验证的、包含100对英语与索马里语有害意图提示的基准数据集）上进行了评估。Llama-3.1-8B-Instruct、Gemma-2-9B-Instruct、Qwen-2.5-7B-Instruct及Aya-23-8B均以温度参数0及相同的英语“有用、无害且诚实”系统提示词本地运行。固定的Claude Sonnet快照版本（claude-sonnet-4-5-20250929）将每个响应分类为拒绝、合规或模糊；母语作者对80行分层样本进行抽查。我们发现所有四个模型均存在显著的英语至索马里语拒绝响应差距：Llama-3.1-8B（0.90；95%自助法置信区间[0.85, 0.96]）、Aya-23-8B（0.75 [0.67, 0.83]）、Qwen-2.5-7B（0.69 [0.59, 0.78]）及Gemma-2-9B（0.38 [0.27, 0.49]）。对于三个模型而言，索马里语非拒绝模式的主要表现形式并非流畅的有害合规响应，而是模糊输出：包括空响应、错误语言或不相干生成结果。母语验证抽查在80个抽样行中与判断器达到100%一致性（Cohen's kappa = 1.00）。我们仅报告聚合拒绝率、类别差距及可靠性统计数据；原始模型生成本地保留，未予发布。