Data as of Jul 25, 2026 · Based on 2,718,867 AI responses across 9,511 prompts · See how Parse measures this
Kakugo is a project that distills knowledge from large language models into smaller models for low-resource languages, providing papers, models, and datasets. It offers data generation and training scripts that rely on a teacher model (GPT-OSS 120B by default) or a Together AI account, and requires Python dependencies such as pandas, datasets, transformers, and numpy. Users configure the target language (e.g., Javanese) via LANG_NAME and run the data-preparation process with scripts like run_prompt_gen.py.
Parse Score