How is just-in-time reinforcement learning being applied to large language models for continuous adaptation?Answer not yet generated.