Text this: Interpretable Merging-Vehicle Acceleration Prediction via a Decoder-Only Transformer and Attention Map Clustering