Hot: transformer-based-large-language-model