With no explanation, label text_A→text_B with either "not-entailed" or "entailed".
text_A: someone perceived a particular person to do a particular thing .
text_B: that person might or might not have done that thing .
entailed.