Scaling mT5 Model Size and Multilingual Intermediate-Task Training Effects on XTREME-R Zero-Shot Transfer

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: How does scaling model size from mT5-small to mT5-large affect the relative gain of multilingual intermediate-task training versus English-only intermediate training on zero-shot transfer accuracy across diverse language families in XTREME-R?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.3/10.