34 5 months ago

This model was fine-tuned on a Turkish corpus consisting of 21,399 Wikipedia articles and approximately 5,000 multiple-choice questions sourced from the alibayram/turkish_mmlu dataset.