Saxo/Linkbricks-Horizon-AI-Korean-Avengers-V3-27B
Saxo/Linkbricks-Horizon-AI-Korean-Avengers-V3-27B is a 27 billion parameter Korean-reinforced language model developed by Yunsung Ji (Saxo) at Linkbricks. Fine-tuned from Saxo/Linkbricks-Horizon-AI-Korean-Avengers-V2-27B, it specializes in high-dimensional analysis of customer reviews and social posts, coding, writing, mathematics, and logical reasoning. The model was trained on 50 million Korean news corpus data, incorporating Korean-Chinese-English-Japanese cross-training data and logic judgment data to enhance multilingual and complex problem-solving capabilities.
Loading preview...
Model Overview
Saxo/Linkbricks-Horizon-AI-Korean-Avengers-V3-27B is a 27 billion parameter Korean-reinforced language model developed by Yunsung Ji (Saxo), a data scientist at Linkbricks. This model is an advanced iteration, fine-tuned from the Saxo/Linkbricks-Horizon-AI-Korean-Avengers-V2-27B base model using 8 H100-80G GPUs, with approximately 38% of its parameters undergoing Korean SFT->DPO training.
Key Capabilities
- Multilingual Proficiency: Enhanced with Korean-Chinese-English-Japanese cross-training data, making it suitable for tasks requiring understanding across these languages.
- Advanced Text Analysis: Excels in high-dimensional analysis of customer reviews and social posts.
- Complex Problem Solving: Strengthened in areas such as coding, creative writing, mathematics, and logical reasoning, trained with specific logic judgment datasets.
- Korean Language Focus: Utilizes a 50 million Korean news corpus for robust Korean language understanding and generation.
Good For
- Applications requiring deep analysis of Korean customer feedback and social media content.
- Developing multilingual applications that need to process or generate text in Korean, Chinese, English, and Japanese.
- Tasks involving complex logical reasoning, mathematical problem-solving, and code generation.
- Creative writing and content generation in Korean and other supported languages.