Bloom Filter Partition Pruning

I have basic understanding of the Bloom filter and the mechanism of the filtering the keys
and I was convinced, that the same is true for the Bloom filter partition pruning. But recently
after investigating a query on a hash partitioned table I'm somehow confused and need some clarification.

Let me illustrate it on a simple example

TABLEA is hash partitioned on the join column TABLEB is a simple table

The query ...

SELECT A.* FROM TABLEA A
WHERE ID IN (SELECT ID FROM TABLEB) ... leads to the following execution plan:

| Id  | Operation                   | Name    | Rows  | Bytes | Cost (%CPU)|
Time     | Pstart| Pstop |
----------------------------------------------------------------------------
---------------------------


|   0 | SELECT STATEMENT            |         | 24242 |  1088K|   460   (1)|


00:00:01 |       |       |
|*  1 |  HASH JOIN RIGHT SEMI       |         | 24242 |  1088K|   460   (1)|
00:00:01 |       |       |


|   2 |   PART JOIN FILTER CREATE   | :BF0000 |     2 |    44 |     3   (0)|


00:00:01 |       |       |
|*  3 |    TABLE ACCESS FULL        | TABLEB  |     2 |    44 |     3   (0)|
00:00:01 |       |       |


|   4 |   PARTITION HASH JOIN-FILTER|         |   400K|  9375K|   456   (1)|


00:00:01 |:BF0000|:BF0000|


|   5 |    TABLE ACCESS FULL        | TABLEA  |   400K|  9375K|   456   (1)|