Open-Orca/Mistral-7B-OpenOrca

Open-Orca/Mistral-7B-OpenOrca is a 7 billion parameter language model developed by Open-Orca, fine-tuned on the Mistral 7B architecture with a 4096-token context length. It leverages a curated subset of the OpenOrca dataset, which is augmented with GPT-4 data, to achieve class-breaking performance on various benchmarks. This model excels in general language understanding and reasoning tasks, outperforming other 7B and 13B models on the HuggingFace Leaderboard at its release.