Improving Model Safety Behavior with Rule-Based Rewardsopenai.com2 points·zerojames··0 commentsOpen articleSaveView on HN