Pre-training or reinforcement learning

[frozen30]

Do the three agents mentioned in the methodology require no prior training? Can they be used directly with existing models?

[frozen30]

Because I see in Problem Solving that the example provided is to directly call the model's API

[fakerNo1]

Yes, they can be used directly with existing models.