| Property | Value |
|---|---|
| Model Type | Large Language Model (Korean) |
| Parameters | 14 Billion |
| Version | 2.0.3 |
| Author | OLAIR |
| Model URL | Hugging Face |
What is ko-r1-14b-v2.0.3?
Ko-r1-14b-v2.0.3 is a specialized Korean language model designed specifically for reasoning tasks. It represents a scaled-up version of the 7B parameter model, focusing on enhanced reasoning capabilities across multiple domains including mathematics, physics, and puzzle-solving.
Implementation Details
The model has been extensively evaluated using the HAE-RAE Reasoning Challenge (HRC), demonstrating competitive performance against larger models. Despite its smaller size compared to some competitors, it achieves notable results, particularly in mathematics and physics domains.
- Achieves 37.68% average score across HAE-RAE benchmark domains
- Outperforms larger models like Exaone-3.5-32B-Instruct and Qwen2.5-72B-Instruct
- Shows particular strength in mathematical reasoning (50.9% score)
Core Capabilities
- Specialized Korean language understanding
- Advanced reasoning abilities across multiple domains
- Strong performance in mathematics and physics tasks
- Efficient scaling characteristics compared to larger models
Frequently Asked Questions
Q: What makes this model unique?
The model's focus on Korean language reasoning tasks and its ability to compete with much larger models make it unique. It demonstrates that targeted training for reasoning can achieve competitive results without requiring extreme model sizes.
Q: What are the recommended use cases?
The model is particularly well-suited for Korean language applications requiring complex reasoning, including mathematical problem-solving, physics applications, and puzzle-solving tasks. However, users should note its current limitations with certain Korean-related inputs that may cause processing loops.