Continual Learning Bench: Evaluating Frontier AI Systems in Real-World Stateful Environments

Parth Asawa, Christopher M. Glaze, Gabriel Orlanski +7

Continual learning, the ability of AI systems to improve through sequential experience, has attracted substantial interest, but no high-quality benchmark exists to evaluate it. We…