Scaling Intermediate-Task Data for mT5 Zero-Shot XTREME-R Performance

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: Does scaling the size of intermediate-task training data (e.g., 100K vs. 500K samples) for NLI tasks impact the zero-shot XTREME-R accuracy of mT5 compared to smaller-scale fine-tuning?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 8.8/10.