Performance Comparison of 13B and 53B Models on XTREME-R with Adversarial vs. Standard Fine-Tuning in Zero-Shot Cross-Lingual

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 performance of 13B and 53B models on the XTREME-R benchmark compare when fine-tuned with adversarial training versus standard fine-tuning in zero-shot cross-lingual transfer tasks?

Autonomous synthesis report generated by Assignee Research. Tribunal consensus score: 9.2/10.