Hugging Face Overview
Model Overview
Saxo/Linkbricks-Horizon-AI-Korean-Advanced-27B is a 27 billion parameter Korean language model developed by Linkbricks Horizon-AI, led by data scientist Yunsung Ji (Saxo). It is built upon the gemma-2-27b-it base model and has undergone extensive Continued Pre-training (CPT), Supervised Fine-tuning (SFT), and Direct Preference Optimization (DPO) using 8 H100-80G GPUs, focusing on approximately 15% of its total parameters.
Key Capabilities
- Multilingual Processing: Trained with cross-lingual data for Korean, Chinese, English, and Japanese, enabling robust performance across these languages.
- Advanced Reasoning: Enhanced for complex logical problems and mathematical tasks.
- Specialized Analysis: Optimized for high-dimensional analysis of customer reviews and social media posts.
- Creative and Technical Writing: Strong capabilities in general writing and code generation.
- Training Methodology: Utilizes Deepspeed Stage=3, rslora, and BAdam Layer Mode for efficient training.
- Tokenizer: Retains the base model's tokenizer without word expansion.
Good For
This model is particularly well-suited for applications requiring advanced Korean language understanding and generation, especially in domains like:
- Analyzing customer feedback and social media trends.
- Developing AI assistants for coding and content creation.
- Solving complex mathematical and logical reasoning problems in a multilingual context.