Fix `PLBart` doctest (#20527)

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

src/transformers/models/plbart/modeling_plbart.py

@@ -559,7 +559,7 @@ PLBART_GENERATION_EXAMPLE = r"""
     >>> values, predictions = probs.topk(5)
 
     >>> tokenizer.decode(predictions).split()
-    ['same', 'first', 'highest', 'result', 'Fib']
+    ['first', 'same', 'highest', 'result', 'number']
     ```
 """