KAERI-MLP/AtomicGPT-gemma2-9B
AtomicGPT-gemma2-9B is a 9 billion parameter large language model developed by the Korea Atomic Energy Research Institute (KAERI), based on Google's Gemma2-9B architecture. This model is specifically specialized in the nuclear engineering domain, having undergone continual pre-training and instruction tuning on relevant data. It excels at providing precise answers to technical and specialized questions across various nuclear technologies, theories, and terminology, supporting both Korean and English languages.
Loading preview...
AtomicGPT-Gemma2-9B: Nuclear Engineering LLM
AtomicGPT-Gemma2-9B is a 9 billion parameter large language model developed by the Korea Atomic Energy Research Institute (KAERI), specifically tailored for the nuclear engineering domain. Built upon the google/gemma-2-9b base model, it has undergone extensive continual pre-training (CPT) and instruction tuning (IT) to deeply understand nuclear technologies, theories, and terminology.
Key Capabilities
- Domain Expertise: Specializes in nuclear engineering topics including reactor design, radiation shielding, the nuclear fuel cycle, and nuclear safety and regulations.
- Multilingual Support: Processes information and queries in both Korean and English.
- Enhanced Performance: Demonstrates significantly improved performance on nuclear-specific evaluations compared to its base Gemma2-9B model, achieving 40% on multiple-choice questions and 19.72% F1 score on short-answer tasks in a zero-shot setting.
Good For
- Technical Q&A: Providing precise and specialized answers to complex questions within the nuclear engineering field.
- Research and Development: Assisting researchers and engineers with information retrieval and understanding of nuclear concepts.
- Educational Applications: Supporting learning and knowledge dissemination in nuclear science and technology.