Weak-to-strong generalization (GPT-4 Turbo was Pareto-SOTA)

2 pointsposted 6 hours ago
by ronfriedhaber
(forourposterity.com)

No comments yet