Jun 2017
Agents learn a reward signal from human comparisons of behavior.
Preference learning becomes part of the path toward helpful language models.
Explore on Humanity Unfolds
