同样句式，地区名称有时候能识别有时候不能识别

同样的句式，识别出来的结果完全不同，如下：
```
上海市徐汇区委书记
[['上海市徐汇区委', '书记'], ['ORG', 'n']]

上海静安区委书记
[['上海静安区委', '书记'], ['ORG', 'n']]

上海普陀区委书记
[['上海普陀区委', '书记'], ['ORG', 'n']]

上海长宁区委书记
[['上海', '长宁', '区委书记'], ['LOC', 'LOC', 'n']]

上海市虹口区委书记
[['上海市', '虹口', '区委书记'], ['LOC', 'LOC', 'n']]
```

Start by reproducing the five Chinese examples and recording the named-entity recognition output for each sentence. Compare the inconsistent location and organization labels, then identify the relevant recognition entry point or model configuration; done means the issue has a defined, tested behavior for these equivalent patterns.

Written by the indexing model from the issue text.