With no explanation, label text_A→text_B with either "strengthener" or "weakener".
text_A: A man with glasses is starring at the camera. A man in sunglasses is having his picture taken.
text_B: The man is about to start talking.
weakener.