Same answer, different IFEval score

[Zor-X-L]

IFEval1.json
IFEval2.json

for example, in IFEval1.json:

        "19": {
            "prompt": [
                {
                    "role": "HUMAN",
                    "prompt": "make a tweet for playboy's twitter account without using capital letters. Include at least 4 hashtags, starting with '#'"
                }
            ],
            "pred": "\"just got out of the shower and feeling all fresh and ready to take on the world. #playboy #freshstart #newday #selfcare\"",
            "refer": {
                "instruction_id_list": [
                    "change_case:english_lowercase",
                    "keywords:letter_frequency"
                ],
                "key": 1122,
                "kwargs": [
                    {},
                    {
                        "let_frequency": 4,
                        "let_relation": "at least",
                        "letter": "#"
                    }
                ],
                "prompt": "make a tweet for playboy's twitter account without using capital letters. Include at least 4 hashtags, starting with '#'"
            },
            "is_strict_correct": false,
            "is_loose_correct": true,
            "is_correct": false,
            "grade": "loose"
        },

with same prompt and prediction, in IFEval2.json:

        "19": {
            "prompt": [
                {
                    "role": "HUMAN",
                    "prompt": "make a tweet for playboy's twitter account without using capital letters. Include at least 4 hashtags, starting with '#'"
                }
            ],
            "pred": "\"just got out of the shower and feeling all fresh and ready to take on the world. #playboy #freshstart #newday #selfcare\"",
            "refer": {
                "instruction_id_list": [
                    "change_case:english_lowercase",
                    "keywords:letter_frequency"
                ],
                "key": 1122,
                "kwargs": [
                    {},
                    {
                        "let_frequency": 4,
                        "let_relation": "at least",
                        "letter": "#"
                    }
                ],
                "prompt": "make a tweet for playboy's twitter account without using capital letters. Include at least 4 hashtags, starting with '#'"
            },
            "is_strict_correct": false,
            "is_loose_correct": false,
            "is_correct": false,
            "grade": "none"
        },

is_loose_correct and grade is different. And I think the answer is strictly correct by the way.

Any hints on where I should start?