Preference data
Records of which of two or more model responses people judged better.
See it in action
Try preference data hands-on in the lesson Instruction tuning and RLHF, part of the How AI assistants work course.
Open the lesson
Records of which of two or more model responses people judged better.
Try preference data hands-on in the lesson Instruction tuning and RLHF, part of the How AI assistants work course.