SciTulu is a collection of instruction-following language models targeting scientific literature understanding use cases. Starting from the Tulu v2 7B model, SciTulu is trained on a mix of science-specific demonstrations from the SciRIFF dataset, together with general-domain instructions from the Tulu v2 SFT mix. SciTulu 7B achives a 28.1% average improvement over Tulu v2 7B on nine held-out scientific literature understanding tasks. More information can be found in our preprint: SciRIFF: A Resource to Enhance Language Model Instruction-Following over Scientific Literature. Training and evaluation code for SciTulu is available in our GitHub repository: https://github.com/allenai/SciRIFF. See the Tulu model card for more information on potential risks, biases, and limitations.
Модальности:
Генерация текста
Задача: Генерация текста
Автор: allenai
Теги: llama, en, text-generation-inference, endpoints_compatible
Лайков: 3 | Загрузок: 140
Описание основано на материалах HuggingFace. Перевод выполнен автоматически.
