Why isn't my trained NER model not able to assign 'O' tag to a sample text?

[Parth-tech]

I had some doubts regarding my trained model and as I am new to spacy, I need some assistance.....
I followed the below steps:

1. I have multiple training files in .TSV format. I directly converted them into .SPACY format. (The TSV file has following the entities: 'GENE', 'DATE', 'QUANTITY', 'DISEASE', 'BODY_PART', 'PERSON', 'O', 'CHEMICAL', 'CELL', 'SPECIES', 'PROCEDURE', 'PLACE', 'PROCESS', 'PHENOMENON', 'VIRUS' and it in the IOB format). I have attached a sample TSV file. (here O means that the word is not important)

2. Then I downloaded a config file, I had chose CPU as Hardware and for Components I chose NER.

3. I trained the model but my testing file was subset of training file so I got unreasonably high score. But the problem is in the Meta.json file I cant see the O tag under NER but I used it in the training file.

10.txt

[polm]

Please copy and paste text as text, not as images.

I trained the model but my testing file was subset of training file so I got unreasonably high score. But the problem is in the Meta.json file I cant see the O tag under NER but I used it in the training file.

The O tag is always present so it's not included in the list, it's automatically added, that's all.

[Parth-tech]

Thank you!