VTI-Data

NLU and NLG datasets developed within the Latvian Language Technology Initiative

1. Alpaca Latvian dataset

   ALPACA-LV is a machine translated Alpaca instruction dataset for Latvian.

2. COPA

   COPA is a machine translated COPA benchmark dataset for Latvian.

3. MMLU

   MMLU is a machine translated MMLU benchmark dataset for Latvian. The sociology_postedited.json file contains a post-edited collection of the first 100 tasks in the sociology subject.