Model Size Scaling Effects on Efficiency-Accuracy Trade-offs in Cross-Lingual Retrieval

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 the scaling of model size (e.g., 100M vs. 7B parameters) affect the trade-off between inference efficiency and zero-shot cross-lingual retrieval accuracy when using intermediate-task training vs. single-task fine-tuning on XTREME-R?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 7.8/10.