Effect of Non-English Intermediate-Task Training on Zero-Shot Cross-Lingual Transfer Performance in Multilingual Models

Intermediate-task training---fine-tuning a pretrained model on an intermediate task before fine-tuning again on the target task---often improves model performance substantially on language understanding tasks in monolingual English settings. We investigate whether English intermediate-task training is still helpful on non-English target tasks. Using nine intermediate language-understanding tasks, we evaluate intermediate-task transfer in a zero-shot cross-lingual setting on the XTREME benchmark. We see large improvements from intermediate training on the BUCC and Tatoeba sentence retrieval tas

Research goal: What is the effect of using non-English intermediate-task training (e.g., Spanish, Chinese) on the zero-shot cross-lingual transfer performance of multilingual models like Llama 3 on TYDI QA, measured by F1 score across languages?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.0/10.