Recent advances in training multilingual language models on large datasets seem to have shown promising results in knowledge transfer across languages and achieve high performance on downstream tasks. However, we question to what extent the current evaluation benchmarks and setups accurately measure zero-shot cross-lingual knowledge transfer. In this work, we challenge the assumption that high zero-shot performance on target tasks reflects high cross-lingual ability by introducing more challenging setups involving instances with multiple languages. Through extensive experiments and analysis, wResearch goal: How does scaling the number of intermediate tasks affect the zero-shot cross-lingual transfer performance of multilingual models on XCOPA and XNLI compared to baseline fine-tuning?Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.9/10.