In the ever-evolving world of artificial intelligence, a new model has emerged that is making waves in the industry. The OpenOrca-Platypus2-13B, a fusion of garage-bAInd/Platypus2-13B and Open-Orca/OpenOrcaxOpenChat-Preview2-13B, has been unveiled. This innovative model is not just a simple combination of its predecessors, but a superior entity that claims to outperform the original Llama-65B model.
The OpenOrca-Platypus2-13B is a testament to the power of collaboration. The creators of the OrcaPlatypus model have joined forces with the Platypus team to develop a model that not only tops the leaderboards but also pushes the boundaries of what is possible in the realm of AI.
OpenOrca Platypus2 13B model
The Language Model Evaluation Harness was employed to run the benchmark tests, using the same version as the HuggingFace LLM Leaderboard. The results were nothing short of impressive. The OpenOrca-Platypus2-13B model demonstrated a performance that was 112% of the base Preview2 model on AGI Eval, averaging 0.463. A significant portion of this boost can be attributed to the remarkable improvement in LSAT Logical Reasoning performance. Check out the video kindly created by the Prompt Engineering YouTube channel below to learn more about what you can expect from this new hybrid model.
Other articles you might find interesting on Llama 2 :
The OpenOrca-Platypus2-13B model is an auto-regressive language model based on the Llama 2 transformer architecture. It was trained by Cole Hunter & Ariel Lee for Platypus2-13B and by Open-Orca for OpenOrcaxOpenChat-Preview2-13B. The model operates in English and is licensed under a Non-Commercial Creative Commons license (CC BY-NC-4.0) for Platypus2-13B base weights and a Llama 2 Commercial license for OpenOrcaxOpenChat-Preview2-13B base weights.
The Open ARCA Platypus 2 13B model, a blend of Open ARCA and Platypus 2 models, has garnered attention due to its stellar performance. The authors assert that this 13 billion parameter model can surpass the original 65 billion parameter model, Lama 65B, on specific datasets. The model was trained on a STEM and logic-based dataset used for the original Platypus 213B model, and the Open ARCA dataset derived using GPT4.
The model excels in logical problems and topics related to science, technology, and engineering. It can generate coherent and well-written text, such as a letter to the CEO of OpenAI, and even programming code. However, users should be aware that the model’s performance on benchmark datasets may not reflect its performance on specific applications, and there may be potential data leakage between training and test sets.
The rapid innovation in the open-source large language model space is both exciting and promising for future developments. The OpenOrca-Platypus2-13B model is a shining example of this innovation, setting a new standard for AI models.
Filed Under: Technology News, Top News
Latest togetherbe Deals
Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, togetherbe may earn an affiliate commission. Learn about our Disclosure Policy.