With no explanation, label text_A→text_B with either "strengthener" or "weakener".
text_A: A boy holding a tube and smiling at the camera. A boy is posing for the camera
text_B: The boy doesn't know who is holding the camera
weakener.