Bilingual vs. English-Only Intermediate-Task Training Efficiency in XTREME-R Benchmarks Across Model Scales

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 intermediate-task training with bilingual tasks (e.g., English + one other language) compare to English-only intermediate-task training in terms of inference efficiency (tokens per second) on XTREME-R benchmark tasks when scaling model size from 7B to 70B parameters?

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