{
  "rule_ja": "NFKC→casefold→空白削除。日本語 20 字（ストライド 10）、英語 40 字（ストライド 20）。12 字未満は対象外。5 設問以上で共有される n-gram は除外",
  "patterns_kept": 911762,
  "patterns_dropped_shared": 16316,
  "dropped_examples": [
    [
      "contradiction",
      1647
    ],
    [
      "forwhichofthesetwoscenariosdoesthemainch",
      895
    ],
    [
      "enariosdoesthemaincharacter(whousesi/me/",
      895
    ],
    [
      "aracter(whousesi/me/my)dosomethingclearl",
      895
    ],
    [
      "my)dosomethingclearlymorallywrong,accord",
      895
    ],
    [
      "ymorallywrong,accordingtoordinarymoralst",
      895
    ],
    [
      "ingtoordinarymoralstandardsintheusasof20",
      895
    ],
    [
      "ong,c.notwrong,wrong,d.notwrong,notwrong",
      895
    ],
    [
      "thisquestionreferstothefollowinginformat",
      606
    ],
    [
      "次の2つのシナリオのうち、主人公(一人称",
      461
    ],
    [
      "うち、主人公(一人称を使う人や話者)が2",
      461
    ],
    [
      "を使う人や話者)が2020年の米国におけ",
      461
    ],
    [
      "020年の米国における通常の道徳的基準に",
      461
    ],
    [
      "る通常の道徳的基準にもとづき明らかに道徳",
      461
    ],
    [
      "もとづき明らかに道徳的に間違っていること",
      461
    ],
    [
      "的に間違っていることをしているのはどちら",
      461
    ],
    [
      "をしているのはどちらですか?シナリオ1|",
      461
    ],
    [
      "いる,d.間違っていない、間違っていない",
      461
    ],
    [
      "non-entailment",
      360
    ],
    [
      "文章:ヨハネス・グーテンベルク[sep]",
      157
    ],
    [
      ",b.false,false,c.true,false,d.false,true",
      102
    ],
    [
      "、真,b.偽、偽,c.真、偽,d.偽、真",
      80
    ],
    [
      "wrong,wrong,b.wrong,notwrong,c.notwrong,",
      64
    ],
    [
      "notwrong,c.notwrong,wrong,d.notwrong,not",
      64
    ],
    [
      ".wrong,notwrong,c.notwrong,wrong,d.notwr",
      56
    ],
    [
      "問題は以下の内容に関するものである。以下",
      54
    ],
    [
      "wrong,notwrong,c.notwrong,wrong,d.notwro",
      54
    ],
    [
      "ng,wrong,b.wrong,notwrong,c.notwrong,wro",
      53
    ],
    [
      "wrong,c.notwrong,wrong,d.notwrong,notwro",
      53
    ],
    [
      "っている、間違っている,b.間違っている",
      53
    ]
  ],
  "coverage": {
    "llmjpeval_aio": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 2000,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 3966,
        "en": 34
      },
      "ngrams": 9286,
      "skipped_short_fields": 1938
    },
    "llmjpeval_alt-e-to-j": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 1010,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "en": 1010,
        "ja": 1010
      },
      "ngrams": 10558,
      "skipped_short_fields": 0
    },
    "llmjpeval_alt-j-to-e": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 1010,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 1010,
        "en": 1010
      },
      "ngrams": 10558,
      "skipped_short_fields": 0
    },
    "llmjpeval_chabsa": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 322,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 579,
        "en": 65
      },
      "ngrams": 3284,
      "skipped_short_fields": 22
    },
    "llmjpeval_commonsensemoralja": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 3992,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 3992,
        "en": 3992
      },
      "ngrams": 8151,
      "skipped_short_fields": 4106
    },
    "llmjpeval_jamp": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 348,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 348,
        "en": 348
      },
      "ngrams": 1884,
      "skipped_short_fields": 232
    },
    "llmjpeval_janli": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 720,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 720,
        "en": 720
      },
      "ngrams": 2528,
      "skipped_short_fields": 360
    },
    "llmjpeval_jcommonsenseqa": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 1119,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 1119,
        "en": 1119
      },
      "ngrams": 4409,
      "skipped_short_fields": 1119
    },
    "llmjpeval_jemhopqa": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 120,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 184,
        "en": 56
      },
      "ngrams": 311,
      "skipped_short_fields": 118
    },
    "llmjpeval_jhumaneval": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 39,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 58,
        "en": 20
      },
      "ngrams": 3041,
      "skipped_short_fields": 0
    },
    "llmjpeval_jmmlu": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 3541,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 3541,
        "en": 3541
      },
      "ngrams": 45851,
      "skipped_short_fields": 3541
    },
    "llmjpeval_jnli": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 2434,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 2434,
        "en": 2434
      },
      "ngrams": 10398,
      "skipped_short_fields": 1700
    },
    "llmjpeval_jsem": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 1584,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 1584,
        "en": 1584
      },
      "ngrams": 6255,
      "skipped_short_fields": 1584
    },
    "llmjpeval_jsick": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 4927,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 4927,
        "en": 4927
      },
      "ngrams": 19150,
      "skipped_short_fields": 4130
    },
    "llmjpeval_jsquad": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 4442,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 8253,
        "en": 631
      },
      "ngrams": 92451,
      "skipped_short_fields": 4004
    },
    "llmjpeval_jsts": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 1457,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 2914
      },
      "ngrams": 6827,
      "skipped_short_fields": 1457
    },
    "llmjpeval_kuci": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 10291,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 10291,
        "en": 10291
      },
      "ngrams": 76585,
      "skipped_short_fields": 10291
    },
    "llmjpeval_mawps": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 500,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 516,
        "en": 484
      },
      "ngrams": 3126,
      "skipped_short_fields": 492
    },
    "llmjpeval_mbpp": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 498,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 842,
        "en": 154
      },
      "ngrams": 16607,
      "skipped_short_fields": 0
    },
    "llmjpeval_mgsm": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 250,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 250,
        "en": 250
      },
      "ngrams": 2670,
      "skipped_short_fields": 250
    },
    "llmjpeval_mmlu_en": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 14042,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "en": 27972,
        "ja": 112
      },
      "ngrams": 263237,
      "skipped_short_fields": 14042
    },
    "llmjpeval_mmmlu": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 7006,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 7005,
        "en": 7007
      },
      "ngrams": 133389,
      "skipped_short_fields": 7006
    },
    "llmjpeval_niilc": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 198,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 381,
        "en": 15
      },
      "ngrams": 307,
      "skipped_short_fields": 216
    },
    "llmjpeval_wiki_coreference": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 200,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 399,
        "en": 1
      },
      "ngrams": 2186,
      "skipped_short_fields": 41
    },
    "llmjpeval_wiki_dependency": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 200,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 400
      },
      "ngrams": 5082,
      "skipped_short_fields": 0
    },
    "llmjpeval_wiki_ner": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 113,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 226
      },
      "ngrams": 2081,
      "skipped_short_fields": 16
    },
    "llmjpeval_wiki_pas": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 199,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 397,
        "en": 1
      },
      "ngrams": 3860,
      "skipped_short_fields": 5
    },
    "llmjpeval_wiki_reading": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 200,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 400
      },
      "ngrams": 3972,
      "skipped_short_fields": 0
    },
    "llmjpeval_wikicorpus-e-to-j": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 3588,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "en": 3571,
        "ja": 3605
      },
      "ngrams": 45569,
      "skipped_short_fields": 311
    },
    "llmjpeval_wikicorpus-j-to-e": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 3588,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 3605,
        "en": 3571
      },
      "ngrams": 45569,
      "skipped_short_fields": 311
    },
    "llmjpeval_xlsum_ja": {
      "source": "zenless-lab/llm-jp-eval（llm-jp-eval 形式ミラー）",
      "split": "test",
      "examples": 677,
      "fields": [
        "input",
        "output"
      ],
      "lang": {
        "ja": 1354
      },
      "ngrams": 90140,
      "skipped_short_fields": 0
    },
    "jglue_jsquad": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 4442,
      "fields": [
        "answer",
        "context",
        "question"
      ],
      "lang": {
        "ja": 19878,
        "en": 1859
      },
      "ngrams": 85151,
      "skipped_short_fields": 11901
    },
    "jglue_jcommonsenseqa": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 1119,
      "fields": [
        "choice",
        "question"
      ],
      "lang": {
        "ja": 6710,
        "en": 4
      },
      "ngrams": 1112,
      "skipped_short_fields": 5952
    },
    "jglue_jnli": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 2434,
      "fields": [
        "sentence1",
        "sentence2"
      ],
      "lang": {
        "ja": 4868
      },
      "ngrams": 8119,
      "skipped_short_fields": 38
    },
    "jglue_jsts": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 1457,
      "fields": [
        "sentence1",
        "sentence2"
      ],
      "lang": {
        "ja": 2914
      },
      "ngrams": 5034,
      "skipped_short_fields": 13
    },
    "jglue_marc_ja": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 5654,
      "fields": [
        "sentence"
      ],
      "lang": {
        "ja": 5653,
        "en": 1
      },
      "ngrams": 97768,
      "skipped_short_fields": 1
    },
    "jglue_jcola": {
      "source": "shunk031/JGLUE（parquet 変換）",
      "split": "validation",
      "examples": 865,
      "fields": [
        "sentence"
      ],
      "lang": {
        "ja": 865
      },
      "ngrams": 950,
      "skipped_short_fields": 180
    },
    "jmmlu": {
      "source": "nlp-waseda/JMMLU（JMMLU.zip）",
      "split": "test",
      "examples": 7536,
      "fields": [
        "choice",
        "question"
      ],
      "lang": {
        "ja": 34839,
        "en": 2841
      },
      "ngrams": 69318,
      "skipped_short_fields": 19149
    },
    "elyza_tasks_100": {
      "source": "elyza/ELYZA-tasks-100（test.csv）",
      "split": "test",
      "examples": 100,
      "fields": [
        "eval_aspect",
        "input",
        "output"
      ],
      "lang": {
        "ja": 300
      },
      "ngrams": 3801,
      "skipped_short_fields": 3
    },
    "japanese_mt_bench": {
      "source": "Stability-AI/FastChat jp-stable question.jsonl",
      "split": "question（60 問: coding・math を除く 6 カテゴリ）",
      "examples": 60,
      "fields": [
        "turn1",
        "turn2"
      ],
      "lang": {
        "ja": 119,
        "en": 1
      },
      "ngrams": 1021,
      "skipped_short_fields": 0
    }
  }
}