Guidance needed: building a Korean dataset for LLaMA fine-tuning

Any advice, resources, or shared experiences would be greatly appreciated. Thank you! 이렇게