regexp_replace return different result compared to Vanilla Spark

### Backend

VL (Velox)

### Bug description

When using regexp, it return different result compared to Vanilla Spark, below is an example:
The issue is related to the processing in '\\,' replace case via Gluten is different than Spark java

create table company_staff stored as parquet as select '执行董事经理' as staff_type_name;

select
regexp_replace(staff_type_name, '、', '\\,')
from company_staff;

### Spark version

Spark 3.2.1

### Spark configurations

_No response_

### System information

_No response_

### Relevant logs

_No response_