N1 2024-12 — audit report

66 questions (問題1–13, the non-listening half). Generated from the OCR and correction audit trails.

OCR confidence

Source: retypeset-digital — quality 4/5: Clean third-party digital retypeset (recall '完成版' edition dated 241225 from jlptmogi.com / Yuuki Bùi / Tôi Yêu Ngoại Ngữ Group), not a scan despite the report: only the cover reproduces the official printed booklet, and all body pages are crisply typeset digital text, fully legible throughout. Heavy but light-gray branding on every page (large circular JLPT MOGI watermark, diagonal name watermark, running header, Vietnamese footer banner, facebook.com/bui.yuuki.9 stamps inside passages) never obscures text. Boxed question numbers 1-66 and 問題 structure are faithfully preserved, including the 問題6 worked example, furigana, and the vertical-text 問題8 (4) passage. Docked one point because it is a retypeset recall version (transcription fidelity not guaranteed against the official booklet) with minor production artifacts: white correction patches over the Q5 area on internal page 1 and footer pagination that skips 6 and 13.

Every section was transcribed twice independently (a full-page pass and a double-resolution half-page pass) and reconciled. Of 13 sections, 6 agreed on all content between the two passes (reconciled automatically, differing only in transcriber notes) and 7 had at least one content difference resolved by a third adjudication pass that re-read the page images. 103 leaf-level differences were examined in total.

Residual OCR risks (25):

• 問題1: The source is a community recall reconstruction, so the kanji/readings themselves may contain recall errors relative to the real exam; transcription reflects the printed page only.
• 問題1: The white correction patches behind q5 mean the group overwrote something there; the visible printed text (すじどう) was transcribed, but what was originally underneath is unknowable from these images.
• 問題2: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題3: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題4: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題5: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題6: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題7: Gloss colon kept half-width per transcribe-exactly-as-printed and because an ASCII colon is not a full-width Japanese-punctuation glyph; a strict normalizer following the convention worked example (which shows a full-width colon) could disagree.
• 問題7: Body furigana over 高 in 高を括っていた reads だか where the 油断 sense expects たか (printed in note 2); transcribed as printed but the small ruby is low-contrast under the watermark.
• 問題8: The exact width of the gloss colon and the trailing space is judged from a low-contrast retype under watermarks; rendered as ASCII ':' + one half-width space per the printed appearance, but a full-width ： reading cannot be fully excluded.
• 問題8: Paragraph breaks in the 縦書き passage (4) are inferred from columns ending short (after 考えていた。 and after 作家＝作曲家。); there are no indents, so the 3-paragraph split carries minor uncertainty.
• 問題8: The (注 2) stray space and the 訳家/演奏家 comma anomalies are transcribed as printed on the assumption they are source recall typos, not OCR artifacts.
• 問題9: Passage (1) isolated gap 子育て中も␣獲物狩り: rendered as a single half-width space (B). The line is fully justified, so it could be justification whitespace rather than a real space; A read it as an artifact. Low-confidence space.
• 問題9: Half- vs full-width parens were judged purely from glyph shape on the scans; if the convention intended to normalize all Japanese-context parens to full-width, passages 1, 2 and the (4) citation would instead use （）. Kept literal per "transcribe exactly what is printed".
• 問題9: Recall-version typos (子とも for 子ども, 長の雌 for 年長の雌, 売り上げを落ちて) are transcribed as printed and not corrected.
• 問題10: no human/agent spot-check was performed for this section (A/B byte-agreement on content was treated as sufficient)
• 問題11: The source is a community recall reconstruction (回忆版); printed wording may itself contain recall typos (e.g. 何のためのこの仕事, possible dropped words in passage B first paragraph). These were transcribed faithfully and not corrected.
• 問題11: Instruction-line spacing was judged to be justification artifacts; if the intent was to preserve literal printed gaps, the chosen normalized form would differ from B.
• 問題12: The full-width 「．」 after 1 in the instruction line is a recall/retype artifact; kept as printed per conventions but ambiguous against a possible faint nakaguro under the watermark.
• 問題12: Whether 一回（注 5） in the body should retain 性 is a source defect, not an OCR uncertainty; transcribed as printed (no 性).
• 問題12: The facebook stamp's exact text (bui.yuuki.9) is read from p29-b but is intentionally omitted from the passage, so it is not verifiable in the final output.
• 問題13: Label-value gap width (full-width vs half-width space) is not pinned down by the conventions; chose full-width to match the ~one-CJK-char printed gaps after CJK labels, but a half-width reading is defensible and the two transcribers genuinely diverged here.
• 問題13: Heading-4 bold extent relies on a subtle weight judgment; high-confidence at double resolution but font rendering of the retype is light.
• 問題13: The 「3.」 box and the rounded メモ box are not representable in minimal HTML; transcribed as plain text per conventions, losing the box framing.
• 問題13: Facebook stamp and watermarks omitted per conventions; not verified beyond visual confirmation of the single bottom-right stamp.

Source-text quirks the OCR transcribed verbatim (printed typos/recall artifacts; the ones judged to be retype errors were corrected in the layer below) — 67 noted at OCR time:

• 問題1: White correction patches are visible over question 5: most clearly behind answer 2「すじどう」and faintly along the latter half of the stem line（筋道を立てて説明した。）. The text printed on top of the patches is legible and was transcribed as printed.
• 問題2: Printed boxed question numbers 7-13 match the expected range; no oddities, injected branding, or suspected recall typos inside the question content on this page.
• 問題3: Question 16 option 1 is printed as めんどうくさ with no ending (verified at high zoom; no trailing く is present). This looks like a recall/retype truncation in the source (めんどうくさく or similar would be expected); transcribed exactly as printed.
• 問題3: Question 15 option 3 かいて is semantically odd as a synonym candidate for ロスして but is clearly printed that way; possibly a recall artifact.
• 問題3: Question 16 underline covers おろそかに (including に); question 15 underline covers ロスして (including て); question 19 underline covers 進呈しました — all verified at high zoom.
• 問題3: Q15 option 3 is printed as 「かいて」 (verified at high zoom: か, not さ). This is likely a recall/retype error for 「さいて」 (割いて), but transcribed as printed.
• 問題3: Q16 option 1 is printed as 「めんどうくさ」 and option 4 as 「いいかげん」, both apparently missing grammatical endings (expected something like 「めんどうくさく」／「いいかげんに」 to parallel options 2 「困難に」 and 3 「遅く」). Transcribed as printed.
• 問題4: q20 option 3 prints 「民家をう安く購入して」 — the stray う after を is in the source (suspected retype typo for 「を安く」); transcribed as printed.
• 問題4: q20 option 4 prints 「就職規則」 — suspected source typo for 就業規則; transcribed as printed.
• 問題4: q21 option 2 prints 「部屋のベテランで植物を養う」 — suspected recall/retype typo for ベランダ; transcribed as printed.
• 問題4: q22 option 2 prints 「見た目は可愛いらしいが」 — suspected typo for 可愛らしい; transcribed as printed.
• 問題4: q23 option 1 prints 「正当な組み立てたのに」 — grammatically odd, suspected recall garble (e.g. 正当に組み立てたのに); transcribed as printed.
• 問題4: q24 option 1 prints 「あの町はにコンビニが」 — stray は before に, suspected typo; transcribed as printed.
• 問題4: q25 option 4 prints 「父と漁民、小さい頃から、海と間柄の深い暮らしをしてきた。」 — phrasing looks garbled in the source (perhaps 父は漁民で…); transcribed as printed.
• 問題4: q20 option 3 prints 「民家をう安く購入して」 with a stray う before 安く — suspected recall/retype typo, transcribed as printed.
• 問題4: q20 option 4 prints 就職規則 (possibly intended 就業規則) — transcribed as printed.
• 問題4: q21 option 2 prints 「部屋のベテランで植物を養うために」 — ベテラン is likely a recall typo for ベランダ, transcribed as printed.
• 問題4: q22 option 2 prints 可愛いらしい (with extra い) and 荒い — transcribed as printed.
• 問題4: q23 option 1 prints 「正当な組み立てたのに」 — ungrammatical な before a verb, likely recall typo, transcribed as printed.
• 問題4: q24 option 1 prints 「あの町はにコンビニが」 with a stray は — transcribed as printed.
• 問題4: q25 options 3 and 4 read awkwardly (「生態の変化、近年の環境の破壊との間柄が指摘される」 / 「父と漁民、小さい頃から、海と間柄の深い暮らしをしてきた」) — likely imperfect recall, transcribed as printed.
• 問題5: q26: the second dialogue line ends 「買う（　）無駄だったよ。」 with past tense だった, which reads oddly in context (expected 無駄だよ); likely a recall typo, transcribed as printed.
• 問題5: q29: the stem reads 「第三者が不正にアクセスされる」 — the particle が with the passive される is grammatically odd (expected 第三者に); likely a recall typo, transcribed as printed.
• 問題5: q31: the options mix the kanji 押さえる (options 1 and 4) and 抑える (options 1, 2, 3) for the same verb; option 3 「抑えるには抑えずに」 looks like a recall typo for the standard 〜には〜たが pattern. All transcribed as printed.
• 問題5: q35: 「報告誌」 verified at high zoom; possibly a recall typo for 広報誌, transcribed as printed.
• 問題5: Q26: the second line ends 「買う（　　）無駄だったよ。」 — past-tense 「だった」 reads oddly after どうせ当たらないんだから (one might expect 「無駄だよ」); transcribed as printed.
• 問題5: Q29: stem reads 「第三者が不正にアクセスされる」 — the particle が with the passive される is grammatically odd (likely a recall/retype slip for 第三者に); transcribed as printed.
• 問題5: Q31: options mix the kanji 押さえる and 抑える — printed as 1「押さえたら抑えたで」, 2「抑えられるものなら」, 3「抑えるには抑えずに」, 4「押さえたいところだが」; the inconsistency (especially within option 1) is likely a retype artifact but was transcribed exactly as printed.
• 問題5: Q31: option 3 「抑えるには抑えずに」 is grammatically odd for the 〜るには〜たが pattern; transcribed as printed.
• 問題5: Q35: the teacher's line says 「青木さん、今度の報告誌のモデルを青木さんにお願いできたらと思うんだけど」 — 青木さん appears twice in the same sentence, which reads oddly (possibly a recall error); transcribed as printed.
• 問題5: Q35: 「報告誌」 is an unusual word (広報誌 would be expected); transcribed as printed.
• 問題6: Q38 answer 1 reads 「堅苦しいものと思われるがちだが」 as printed — 「思われるがちだが」 looks like a recall/retype error for 「思われがちだが」; transcribed exactly as printed.
• 問題6: q38 option 1 is printed as 「堅苦しいものと思われるがちだが」 with an extra る (standard Japanese would be 思われがちだが); transcribed exactly as printed — likely a recall/retype error in the source.
• 問題6: The 問題例/解答のしかた worked-example block on page 9 was skipped per conventions; the worked example itself contains a typo 「正しい分はこうです」 (分 for 文), reinforcing that this is a retyped recall source.
• 問題7: Suspected recall/retype typo transcribed as printed: the paragraph beginning これが一人で… ends 「お疲れさまと入ったであった。」, which does not parse (a noun such as 店 seems dropped after 入った); verified at high zoom that nothing is printed between 「入ったで」 and 「あった」.
• 問題7: Furigana inconsistency in the source: the passage body prints だか over 高 in 高を括っていた, while （注2） on p12 prints たか over the same 高; both transcribed as printed.
• 問題7: Unusual furigana transcribed as printed: 工合 is glossed こうごう in both the passage body and the （注3） gloss; the expected reading for the gloss 具合 would be ぐあい — possibly a retype error.
• 問題7: The （注）glosses are printed with a half-width colon followed by a space (e.g. 「発注する: 注文する」) rather than a full-width ：; kept as printed. The surrounding parentheses are rendered full-width and the digits half-width per conventions.
• 問題8: Passage (3): the second note label is printed as (注 2) with a stray space between 注 and 2 (while the first is (注1) without a space); transcribed as printed.
• 問題8: Passage (4), first paragraph: printed as 「作家は作曲家に訳家は演奏家に、それぞれ…」 with no comma after 作曲家に and 訳家 rather than 翻訳家 — suspected recall/retype error (expected something like 「作家は作曲家に、翻訳家は演奏家に」); transcribed exactly as printed.
• 問題8: Passage (4), third paragraph: 「…翻訳家＝演奏家、楽譜をもとにしていても…」 — the comma after 演奏家 is as printed, though a sentence-ending 。would be expected there; transcribed as printed.
• 問題9: Parenthesis/colon widths are inconsistent across passages and transcribed exactly as printed: passages (1), (2) and the (4) citation use half-width ASCII parens and half-width colons (e.g. 捕食獣(注1), (注 1)捕食獣: ...), while passage (3) uses full-width （）and full-width ： (e.g. マーケティング調査（注1）, （注1）マーケティング調査： 市場調査). The space after each 注-gloss colon is present in the source and kept.
• 問題9: Passage (1): printed 「誰かが子とものそばに残っていれば」 — 子とも (no dakuten) appears to be a retype typo for 子ども; transcribed as printed.
• 問題9: Passage (1): printed 「長の雌がまとめて面倒を見た方が」 — likely missing 年 (question 49 option 2 has 経験豊富な年長の雌); transcribed as printed.
• 問題9: Passage (1) body uses 一か月 (hiragana か) while question 49 option 2 prints 一カ月 (katakana カ); both transcribed as printed.
• 問題9: (中略) appears inline (not on its own line) in passages (1) and (2); printed with full-width parens in (1) and half-width parens in (2); transcribed as printed.
• 問題9: Passage (3): printed 「従来よりも売り上げを落ちてしまっては」 — ungrammatical particle (expected 売り上げが落ちて or 売り上げを落として); transcribed as printed.
• 問題9: Passage (3): printed 「今までのファンがある程度を許容してくれれば」 — the を after ある程度 is unusual; transcribed as printed.
• 問題10: No source citation line and no 注 glosses are printed after the passage (likely omitted by the recall retypesetters).
• 問題10: Question 59 answer 1 prints 「経験に満足するとなく」, almost certainly a recall/retype typo for 「経験に満足することなく」; transcribed exactly as printed.
• 問題10: The passage text mixes 学生たち and 生徒 when referring to the same people (e.g. paragraph 2 vs paragraphs 3-4); transcribed as printed.
• 問題10: Question 59 option 1 prints 「経験に満足するとなく、自ら変化し続けようとすること。」 — 「するとなく」 is a suspected recall/retype typo for 「することなく」; verified at high zoom and transcribed exactly as printed.
• 問題11: Passage A, second paragraph: printed as 「何のためのこの仕事を行い」 — grammatically odd; suspected recall/retype typo for 「何のためにこの仕事を行い」. Transcribed as printed.
• 問題11: Passage A, first paragraph: no comma between 「不足していたり」 and 「仕事に対して」 — verified at high zoom, transcribed as printed.
• 問題11: Passage B, first paragraph: 「指示待ちになっている理由が何かあるのかもしれません」 reads slightly unnatural; possibly words dropped in the recall retype. Transcribed as printed.
• 問題11: Passage B, fourth paragraph: 「上司の思いにそって」 printed in kana (そって), transcribed as printed.
• 問題11: Question 61 stem reads 「部下や後輩の対応のしかたについて」 (no へ before の対応); possibly a retype omission, transcribed as printed.
• 問題12: The printed instruction line reads 「最もよいものを、1．2・3・4から一つ選びなさい。」 with a baseline full-width period after the 1 instead of the usual nakaguro (1・2); transcribed as printed.
• 問題12: Suspected recall typos in paragraph 3: the source prints 「普通的な法則」 and 「科学の普通法則」 where 普遍的な法則／普遍法則 is clearly intended (question 63's options all use 普遍的な法則); transcribed as printed.
• 問題12: Paragraph 3 prints 「一回（注 5）」 with a half-width space before the 5, while the gloss defines 一回性; the source has apparently dropped the 性 (the original presumably reads 一回性（注5）); transcribed as printed.
• 問題12: Inconsistent （注N） spacing in the source: in-passage references are set tight （注1）（注2）（注3）（注4）（注6） except （注 5） which has a half-width space; all six gloss lines print a space, （注 1）…（注 6）, and a half-width space after the full-width colon (e.g. よそよそしい： 無関係である). Both styles kept as printed.
• 問題12: 「論理や価値観に属する話です」 — 論理 is likely a retype error for 倫理 (ethics); transcribed as printed.
• 問題13: The 問題13 instruction says 「ある大学の研究活動助成の案内」 but the document is a ボランティア活動助成金 (volunteer-activity grant) guide — apparent recall/retype mismatch in the source; transcribed as printed.
• 問題13: Source typo in the 交通費 line: 「10,000円を上限としてを支給します」 (extra を before 支給, compare the 宿泊費 line 「上限として実費を支給します」); transcribed exactly as printed.
• 問題13: Group branding injected into the document content: section 7 reads 「緑川大学ボランティアセンター（3号館1階―Yuuki）」 and the e-mail is jlptmogi1@gmail.com (printed underlined, kept as <u>); all transcribed as printed.
• 問題13: Document line C prints a full-width colon directly before the parenthetical: 「ボランティア活動証明書：（受け入れ先の様式も可）」; transcribed as printed.
• 問題13: Heading 4 is entirely bold including the parenthetical 「（A〜C…ダウンロード可）」 (verified against the double-resolution image: the parenthetical matches the bold weight of 申請方法).

Corrections applied (27)

Evidence-based reconstructions of the original exam text. This section is the canonical correction log for the sitting.

• 問4 q20 answers[2]: 「をう安く」 → 「を安く」
  • 「をう安く」is a clear non-word; the unambiguous intended form is 「を安く」(古い民家を安く購入して). The stray う is a typo in the source reconstruction, faithfully OCR'd. Meets the clear-non-word correction bar.
  • evidence: vision-vs-scan input/N1_2024-12/yuuki-bui-exam-retype.pdf p4; no other text witness covers 問題4 distractor sentences
• 問4 q20 answers[2]: 「購入して<u>加工</u>して、民宿に加工した」 → 「購入して、民宿に<u>加工</u>した」
  • Source-reconstruction duplication: the 完成版 reconstruction repeats 加工 twice (購入して加工して、民宿に加工した); the true exam usage distractor uses it once. Remove the extra 加工して; the underlined target moves to the remaining 加工. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Human reconstruction; no genuine booklet scan exists (only the Yuuki 完成版).
• 問4 q23 answers[0]: 「<u>正当</u>な組み立てた」 → 「<u>正当</u>に組み立てた」
  • Source-reconstruction grammar: 正当な + 組み立てた is ungrammatical; the real exam reads 正当に組み立てた. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Human reconstruction; no genuine booklet scan exists.
• 問4 q24 answers[0]: 「あの町はにコンビニ」 → 「あの町にはコンビニ」
  • Source-reconstruction grammar: double particle はに; natural form is あの町には. Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Human reconstruction; no genuine booklet scan exists.
• 問4 q25 answers[3]: 「父と漁民、小さい頃」 → 「父は漁民で、小さい頃」
  • Source-reconstruction grammar: 父と漁民、 reads oddly; the real exam reads 父は漁民で、(間柄 remains the deliberately-misused target word). Per Domenic's human judgement (OCR-confidence audit notes), implemented 2026-06-17.
  • evidence: Human reconstruction; no genuine booklet scan exists.
• 問9 q53 passage: 「売り上げを落ちて」 → 「売り上げが落ちて」
  • 問9(3): 落ちる is intransitive (takes が); 「売り上げを落ちて」is ungrammatical. Source 佐藤卓『大量生産品のデザイン論』reads 「売り上げが落ちて」. Recall/OCR error in the witness-less 完成版. Found via the Google Books source-reconciliation pass (2026-06-17, Domenic-authorized); evidence = verbatim full-text match, see docs/passage-sources.md.
  • evidence: Google Books full-text: 佐藤卓『大量生産品のデザイン論』
• 問9 q53 passage: 「ある程度を許容」 → 「ある程度は許容」
  • 問9(3): 「ファンがある程度は許容してくれれば」— ある程度は is adverbial; 「ある程度を許容」is awkward. Source 佐藤卓『大量生産品のデザイン論』reads は. Found via the Google Books source-reconciliation pass (2026-06-17, Domenic-authorized); evidence = verbatim full-text match, see docs/passage-sources.md.
  • evidence: Google Books full-text: 佐藤卓『大量生産品のデザイン論』
• 問12 q62 passage: 「論理や価値観に属する」 → 「倫理や価値観に属する」
  • 問12: whether an act is 許される is a matter of ethics, so 倫理 not 論理 (論↔倫 recall/OCR confusion). Source 岸田一隆『科学コミュニケーション』reads 倫理. Found via the Google Books source-reconciliation pass (2026-06-17, Domenic-authorized); evidence = verbatim full-text match, see docs/passage-sources.md.
  • evidence: Google Books full-text: 岸田一隆『科学コミュニケーション』
• 問12 q62 passage: 「普通的な法則に到達するまでの道のり」 → 「普遍的な法則に到達するまでの道のり」
  • 普通的な法則 is not the intended word; question 63's options all use 普遍的な法則 (3 occurrences). 普通↔普遍 recall/OCR confusion in the witness-less 完成版.
  • evidence: Internal: question 63 options all read 普遍的な法則 (e.g. '普遍的な法則にたどり着けない', '普遍的な法則にたどり着いたとしても'). Google Books: 岸田一隆『科学コミュニケーション』 snippet '普遍的な法則に到達するまでの道のりはきわめて遠いのです'. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問12 q62 passage: 「科学の普通法則は人生の個別性」 → 「科学の普遍法則は人生の個別性」
  • Same 普通→普遍 confusion as above; the source reads 科学の普遍法則.
  • evidence: Internal: passage uses 普遍 in question 63 options; Google Books snippet '科学の普遍法則は人生の個別性や一回性に対しては全く無力です'. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問12 q62 passage: 「個別性や一回（注 5）に対しては」 → 「個別性や一回性（注 5）に対しては」
  • 性 dropped from 一回性; the （注5）gloss in the same passage reads 「一回性： 一回起こるだけ」, so the note marker must attach to the full word 一回性.
  • evidence: Internal: note 5 gloss in this passage reads '（注 5）一回性： 一回起こるだけ'. Google Books snippet '人生の個別性や一回性に対しては'. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問7 q41 passage: 「工合<rt>こうごう</rt></ruby>（注3）を」 → 「工合<rt>ぐあい</rt></ruby>（注3）を」
  • こうごう is not a reading of 工合 (an old form of 具合, read ぐあい). The （注3）gloss confirms the word is 具合. Recall/retype furigana error.
  • evidence: Dictionary: 工合/具合 reads ぐあい, never こうごう; （注3）glosses it as 具合. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問7 q41 passage: 「（注3）<ruby>工合<rt>こうごう</rt></ruby>: 具合」 → 「（注3）<ruby>工合<rt>ぐあい</rt></ruby>: 具合」
  • Same wrong reading こうごう in the note gloss; 工合 (= 具合) reads ぐあい.
  • evidence: Dictionary: 工合/具合 reads ぐあい, never こうごう. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問8 q48 passage: 「作家は作曲家に訳家は演奏家に」 → 「作家は作曲家に翻訳家は演奏家に」
  • 訳家 is a clear non-word; the same passage uses 翻訳家 repeatedly (翻訳家＝演奏家, 翻訳家の個性や解釈). Recall/retype dropped 翻. Number corrected from the proposed 45 to the actual block (48); comma insertion dropped as unjustified overreach since the parallel clause needs no separating comma.
  • evidence: Internal: passage paragraph 3 reads '翻訳家＝演奏家' and '翻訳家の個性や解釈によって'; 訳家 is a non-word. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q49 passage: 「誰かが子とものそばに残っていれば」 → 「誰かが子どものそばに残っていれば」
  • 子とも is a clear non-word (dropped dakuten); the same passage writes 子ども/子どもたち elsewhere.
  • evidence: Internal: passage uses 子ども/子どもたち repeatedly (e.g. 'それらの子どもたちが集まって'). [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問9 q49 passage: 「長の雌がまとめて面倒を見た方が」 → 「年長の雌がまとめて面倒を見た方が」
  • 長の雌 is not a word; the intended word is 年長の雌, matching question 49 option 1's 経験豊富な年長の雌 and the surrounding text about 壮年期 experienced lionesses. 年 dropped in recall/retype.
  • evidence: Internal: question 49 option reads '経験豊富な年長の雌が育てる'; passage mentions '壮年期になったライオンは経験豊富'. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問10 q59 answers[0]: 「経験に満足するとなく」 → 「経験に満足することなく」
  • 「満足するとなく」 is ungrammatical; the standard form 「満足することなく」 (without being satisfied) is intended. Dropped こ in recall/retype.
  • evidence: Grammar: 〜することなく is the standard form; するとなく is a non-word. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問11 q60 passageA: 「何のためのこの仕事を行い」 → 「何のためにこの仕事を行い」
  • 「何のためのこの仕事を行い」 is ungrammatical; 「何のために」(for what purpose) is required before the verb 行い. の↔に recall/retype slip.
  • evidence: Grammar: 何のために + verb is required; 何のための + verb-clause is ungrammatical. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問6 q38 answers[0]: 「堅苦しいものと思われるがちだが」 → 「堅苦しいものと思われがちだが」
  • 〜がち attaches to the ます-stem, so 思われ + がち = 思われがち; 思われる + がち has an extra る and is ungrammatical. The sentence-ordering fragment must be well-formed.
  • evidence: Grammar: 〜がち takes the masu-stem (思われがち), not the dictionary form. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q20 answers[3]: 「就職規則の一部を」 → 「就業規則の一部を」
  • 就職規則 is a non-word; the established term is 就業規則 (work rules). The 問題4 deliberate misuse concerns the target word 加工, not the surrounding noun, which should be natural Japanese. 職↔業 recall/retype confusion.
  • evidence: Standard term is 就業規則; 就職規則 is non-standard. The distractor mechanism (misuse of 加工) is untouched. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q22 answers[1]: 「見た目は可愛いらしいが」 → 「見た目は可愛らしいが」
  • 可愛いらしい is non-standard; the adjective is 可愛らしい (かわいらしい). The 問題4 deliberate misuse concerns the target word 資質, not this adjective, which should be natural. Recall/retype error.
  • evidence: Standard form is 可愛らしい (かわいらしい); 可愛いらしい is non-standard. The distractor mechanism (misuse of 資質) is untouched. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問5 q29 question: 「第三者が不正にアクセスされる」 → 「第三者に不正にアクセスされる」
  • With the passive される, the agent doing the unauthorized access takes に: 「第三者に不正にアクセスされる恐れがある」. が with される makes the third party the patient, which is the wrong meaning. が↔に recall/retype slip.
  • evidence: Grammar/semantics: passive される takes the agent with に; the third party is the one accessing, so 第三者に...される. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問3 q16 answers[0]: 「めんどうくさ」 → 「めんどうくさく」
  • めんどうくさ is a truncated non-word (the adjective is めんどうくさい). Filling 「確認が＿＿なっていた」 it parallels options 困難に and 遅く, so the adverbial めんどうくさく is expected. Dropped ending in recall/retype.
  • evidence: Truncation (めんどうくさい→adverbial めんどうくさく); siblings 困難に/遅く attach to なっていた. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問3 q16 answers[3]: 「いいかげん」 → 「いいかげんに」
  • Like the sibling options, this fills 「確認が＿＿なっていた」 and must attach to なっていた, so いいかげんに is expected, parallel to 困難に and 遅く. Dropped に in recall/retype.
  • evidence: Parallel structure: options attach to なっていた (困難に/遅く/いいかげんに); いいかげん is a な-adjective requiring に before なる. [correct-retypes workflow (propose + adversarial verify), 2026-06-17]
• 問4 q21 answers[1]: 「部屋のベテランで植物を<u>養う</u>ためにお店で鉢と土を買った。」 → 「部屋のベランダで植物を<u>養う</u>ためにお店で鉢と土を買った。」
  • 「部屋のベテランで植物を養う」 is nonsensical; ベテラン (veteran) is a retype garble for ベランダ (balcony), where one would place a pot and soil to grow plants. The misuse being tested in this 問題4 distractor is 養う applied to plants, not the word ベランダ, so correcting the retype typo does not damage the intended distractor.
  • evidence: Category typo, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問13 q65 passage: 「10,000円を上限としてを支給します。」 → 「10,000円を上限として支給します。」
  • Stray を before 支給 makes the clause ungrammatical (no object precedes 支給します). The parallel 宿泊費 line reads 「上限として実費を支給します」; the 交通費 line should read 「上限として支給します」 without the extra を.
  • evidence: Category added-char, confidence high. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]
• 問5 q35 question: 「今度の報告誌のモデルを青木さん」 → 「今度の広報誌のモデルを青木さん」
  • 報告誌 is not a standard word and a 「モデル」 (cover/photo model) belongs to a 広報誌 (publicity magazine), not a 報告誌. 広報誌 is the intended real-world referent; 報告 is a retype substitution for 広報.
  • evidence: Category homophone, confidence medium. [retype-corrections-pass2 (propose + adversarial verify), 2026-06-20]

Flagged unresolved text issues (1)

Ambiguous source-text issues left unchanged rather than guessed.

• 問4 q22 answers[0] — disproven-earlier-claim
  • An earlier underline-review pass claimed our 資質 should be 遮断 (q22) and 正当 should be 要請 (q23). VERIFIED FALSE against the image scan (p4): the source plainly prints 資質 and 正当, matching our text exactly. No change needed. Recorded here so the claim is not re-raised.