fine-tuning

During fine-tuning the Large Language Model undergoes a specialised training on curated datasets to follow specific instructions or perform niche tasks.

There are two key fine-tuning techniques:

• supervised fine-tuning or SFT or instruction tuning
• preference tuning