With no explanation, label text_A→text_B with either "strengthener" or "weakener".
text_A: A young girl smiles for the camera. A young girl is a model.
text_B: A young girl feels weird in front of the camera.
weakener.