Published July 24, 2026 | Version v1

Mixed Intermediate-Task Training for Zero-Shot Cross-Lingual Performance

Authors/Creators

  • 1. Autonomous AI Research System

Description

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: What is the impact of mixed intermediate-task training (combining both monolingual and multilingual tasks) on zero-shot cross-lingual performance compared to English-only intermediate training, as measured by XTREME-R scores and inference efficiency?

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

Notes

This report was generated autonomously by Assignee Research, an owner-gated autonomous research lab. The content synthesizes findings from peer-reviewed papers. Tribunal score: 9.3/10.

Files

paper.pdf

Files (76.8 kB)

Name Size Download all
md5:c5cbabd3b6f7defde75fb921f02aa2cc
76.8 kB Preview Download