English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Instruction tuning has remarkably advanced large language models (LLMs) inunderstanding and responding to diverse human instructions. Despite the successin high-resource languages, its application in lower-resource ones faceschallenges due to the imbalanced foundational abilities of LLMs acrossdifferent languages, stemming from the uneven language distribution in theirpre-training data. To tackle this issue, we propose pivot language guidedgeneration (PLUG), an approach that utilizes a high-resource language,primarily English, as the pivot to enhance instruction tuning in lower-resourcelanguages. It trains the model to first process instructions in the pivotlanguage, and then produce responses in the target language. To evaluate ourapproach, we introduce a benchmark, X-AlpacaEval, of instructions in 4languages (Chinese, Korean, Italian, and Spanish), each annotated byprofessional translators. Our approach demonstrates a significant improvementin the instruction-following abilities of LLMs by 29% on average, compared todirectly responding in the target language alone. Further experiments validatethe versatility of our approach by employing alternative pivot languages beyondEnglish to assist languages where LLMs exhibit lower proficiency.

PLUG: Leveraging Pivot Language in Cross-Lingual Instruction Tuning