Building Instruction-Tuning Datasets from Human-Written Instructions with Open-Weight Large Language Models

Youmi Ma, Sakae Mizuki, Kazuki Fujii +12

Instruction tuning is crucial for enabling Large Language Models (LLMs) to solve real-world tasks. Prior work has shown the effectiveness of instruction-tuning data synthesized sol…