Multilingual contextual embeddings have demonstrated state-of-the-art performance in zero-shot cross-lingual transfer learning, where multilingual BERT is fine-tuned on one source language and evaluated on a different target language. However, published results for mBERT zero-shot accuracy vary as much as 17 points on the MLDoc classification task across four papers. We show that the standard practice of using English dev accuracy for model selection in the zero-shot setting makes it difficult to obtain reproducible results on the MLDoc and XNLI tasks. English dev accuracy is often uncorrelateResearch goal: How does using target-language-specific development sets for model selection impact the zero-shot cross-lingual performance of multilingual BERT on the XNLI dataset compared to other multilingual language models like XLM-R or mT5, as measured by macro-averaged F1 scores across all languages?Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.8/10.