ConceptSFT / 监督式指令微调
Supervised Fine-Tuning
A fine-tuning method that trains LLMs on annotated instruction-response pairs so they follow instructions and complete tasks. It is a common approach in LLM alignment training.
A fine-tuning method that trains LLMs on annotated instruction-response pairs so they follow instructions and complete tasks. It is a common approach in LLM alignment training.