cturan/lamba-turkish-sft

--- license: odc-by language: - tr task_categories: - question-answering - translation - summarization - text-generation - text-classification --- # Lamba Turkish SFT Dataset This is a Turkish Supervised Fine-Tuning (SFT) dataset. The topic distribution is largely aligned with the Turkish High School (Lise) curriculum. It contains a wide variety of examples designed to improve model capabilities in: - Instruction following - Summarization (Özetleme) - Information extraction (Bilgi çıkarma) - General problem solving I hope this dataset will be beneficial to the open-source and AI community. ## Disclaimer Since the vast majority of this dataset is generated synthetically via language models, absolute factual accuracy cannot be guaranteed. Examples may contain hallucinations, errors, or inaccuracies. Please keep this in mind while using this dataset. ## Credits & License A portion of this dataset (around 10k Turkish interactions) was filtered from the [allenai/WildChat-4.8M](https://huggingface.co/datasets/allenai/WildChat-4.8M) dataset. The dataset is distributed under the same license terms as the constituent subsets (abiding by the licensing constraints of allenai/WildChat).