Rank Response to align Human Feedback
efficiently align language model output probabilities with human preferences as robust as fine-tuning and it only needs 1 to 2 models during tuning
RRHF
Creator
Creator
Seonglae ChoCreated
Created
2023 Jul 15 17:5Editor
Editor
Seonglae ChoEdited
Edited
2023 Sep 9 17:10Refs
Refs