Using DeepSeek R1 we can see his train of thought, which is cut off at some point. But never do we get the actual response which should replace the train of thought.