Advanced
Artificial Intelligence & Machine Learning
Q87 / 100
What is Constitutional AI (CAI) and how does it improve alignment?
Correct! Well done.
Incorrect.
The correct answer is B) Anthropic's technique for self-critiquing model outputs against a list of principles and using the critiques to generate better responses, reducing dependence on human labelers
B
Correct Answer
Anthropic's technique for self-critiquing model outputs against a list of principles and using the critiques to generate better responses, reducing dependence on human labelers
Explanation
CAI (Bai et al., 2022): SL-CAI critiques and revises its own outputs against a "constitution" of principles. RL-CAI uses AI feedback instead of human feedback. Scales alignment without proportional human labeling cost.
Progress
87/100