Marimonald/ojisan-translation-corpus-ja

Ojisan Translation Corpus (Japanese) 日本語のドキュメントはこちら The Ojisan Translation Corpus (ojisan-translation-corpus-ja) is a curated, high-quality Japanese dataset designed for style-transfer and preference alignment into the culturally unique "Ojisan-dialect" (おじさん構文 / Ojisan-koubun). This dataset powers the fine-tuning of Ojisan-translator-v1-Qwen-3.5-9B, providing structured subsets for Supervised Fine-Tuning (SFT) and Kahneman-Tversky Optimization (KTO). Dataset… See the full description on the dataset page: https://huggingface.co/datasets/Marimonald/ojisan-translation-corpus-ja.