microsoft/Orca-2-13b

microsoft/Orca-2-13b is a 13 billion parameter language model, fine-tuned from LLAMA-2, specifically designed for research into enhancing small language models' reasoning capabilities. It excels in tasks such as reasoning over user-given data, reading comprehension, math problem-solving, and text summarization, primarily through advanced prompting and synthetic data training. This model is intended to demonstrate how complex workflows can teach SLMs new capabilities, particularly in reasoning.