8015fdbb79
Replace nearest-neighbour personal routing with a frozen class-balanced linear head measured on historical, stratified, cross-validation, holdout, and fresh challenge gates (V-702). Close the four repair handoff holes, preserve nested clarification flows, and route Russian possession statements through structural grammar rather than lexical exceptions (V-573). Owner explicitly requested direct commits to master.
522 lines
16 KiB
JSON
522 lines
16 KiB
JSON
{
|
||
"schema_version": 1,
|
||
"name": "personal_boundary_v1",
|
||
"notes": [
|
||
"Held-out personal-boundary fixture for V-702. Cases are not production seeds and the fixture test enforces that after case folding and punctuation removal.",
|
||
"The matrix is balanced: three cases for every stratum × language × expected side cell (6 × 2 × 2 × 3 = 72).",
|
||
"Personal means that only the owner's local data can answer. World means that an outside knowledge source can answer even when the wording uses first person or possession.",
|
||
"Public subjects are deliberately varied. No case repeats the Baikal regression or changes only its place name."
|
||
],
|
||
"cases": [
|
||
{
|
||
"id": "pb-ru-remembered-personal-01",
|
||
"utterance": "что я раньше говорил насчёт бессонницы?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-ru-remembered-personal-02",
|
||
"utterance": "помнишь, какие причины увольнения я называл?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-ru-remembered-personal-03",
|
||
"utterance": "упоминал ли я, куда хочу переехать?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-ru-remembered-world-01",
|
||
"utterance": "что Чехов говорил о краткости?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-ru-remembered-world-02",
|
||
"utterance": "какие причины кризиса называли экономисты?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-ru-remembered-world-03",
|
||
"utterance": "что свидетели рассказывали об извержении Кракатау?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-personal-01",
|
||
"utterance": "what was it I told you about changing careers?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-personal-02",
|
||
"utterance": "have I ever mentioned where I studied?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-personal-03",
|
||
"utterance": "do you remember which camera I said I preferred?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-world-01",
|
||
"utterance": "what did Marie Curie write about radium?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-world-02",
|
||
"utterance": "which causes of inflation do economists usually mention?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
{
|
||
"id": "pb-en-remembered-world-03",
|
||
"utterance": "what did the Apollo astronauts report about lunar dust?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "remembered_speech"
|
||
},
|
||
|
||
{
|
||
"id": "pb-ru-possession-personal-01",
|
||
"utterance": "какой номер у моего страхового полиса?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-ru-possession-personal-02",
|
||
"utterance": "где я оставил свои запасные ключи?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-ru-possession-personal-03",
|
||
"utterance": "до какого числа действует мой абонемент?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-ru-possession-world-01",
|
||
"utterance": "как убрать царапину с моего стола?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-ru-possession-world-02",
|
||
"utterance": "почему у меня запотевают окна зимой?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-ru-possession-world-03",
|
||
"utterance": "чем зарядить мой телефон в поездке?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-personal-01",
|
||
"utterance": "when does my library card expire?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-personal-02",
|
||
"utterance": "where did I put my passport copy?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-personal-03",
|
||
"utterance": "what size are my hiking boots?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-world-01",
|
||
"utterance": "how can I descale my kettle safely?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-world-02",
|
||
"utterance": "why does my laptop fan get loud under load?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
{
|
||
"id": "pb-en-possession-world-03",
|
||
"utterance": "which adapter should I use for my phone abroad?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "possession"
|
||
},
|
||
|
||
{
|
||
"id": "pb-ru-narrative-personal-01",
|
||
"utterance": "напомни историю о том, как я познакомился с Антоном",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-ru-narrative-personal-02",
|
||
"utterance": "расскажи, что со мной случилось в первый день на новой работе",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-ru-narrative-personal-03",
|
||
"utterance": "восстанови по моим заметкам историю поездки в Казань",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-ru-narrative-world-01",
|
||
"utterance": "опиши восхождение на Эверест",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-ru-narrative-world-02",
|
||
"utterance": "расскажи историю создания языка Rust",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-ru-narrative-world-03",
|
||
"utterance": "объясни, как возникли кольца Сатурна",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-personal-01",
|
||
"utterance": "retell the story of how I met Lena from what I told you",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-personal-02",
|
||
"utterance": "walk me through what happened on my first day at university",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-personal-03",
|
||
"utterance": "reconstruct my Prague trip from my notes",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-world-01",
|
||
"utterance": "tell me the story of the first Moon landing",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-world-02",
|
||
"utterance": "describe how the printing press spread through Europe",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
{
|
||
"id": "pb-en-narrative-world-03",
|
||
"utterance": "explain how the Panama Canal was built",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "narrative"
|
||
},
|
||
|
||
{
|
||
"id": "pb-ru-preamble-personal-01",
|
||
"utterance": "если помнишь наш разговор, что я решил насчёт переезда?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-ru-preamble-personal-02",
|
||
"utterance": "как я уже упоминал, когда мне продлевать страховку?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-ru-preamble-personal-03",
|
||
"utterance": "возвращаясь к тому, что я рассказывал, какую модель велосипеда я выбрал?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-ru-preamble-world-01",
|
||
"utterance": "как я уже говорил, почему самолёты оставляют белый след?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-ru-preamble-world-02",
|
||
"utterance": "возвращаясь к моему вопросу, из чего состоит базальт?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-ru-preamble-world-03",
|
||
"utterance": "я, возможно, повторяюсь, но когда построили Колизей?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-personal-01",
|
||
"utterance": "as I mentioned earlier, which dentist did I choose?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-personal-02",
|
||
"utterance": "coming back to what I told you, when am I taking leave?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-personal-03",
|
||
"utterance": "I may have said this already, which Linux distro did I settle on?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-world-01",
|
||
"utterance": "as I was saying, why do tides happen?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-world-02",
|
||
"utterance": "coming back to my question, how are auroras formed?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
{
|
||
"id": "pb-en-preamble-world-03",
|
||
"utterance": "I may be repeating myself, when was Machu Picchu built?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "first_person_preamble"
|
||
},
|
||
|
||
{
|
||
"id": "pb-ru-advice-personal-01",
|
||
"utterance": "что из моих дел нужно закончить до пятницы?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-ru-advice-personal-02",
|
||
"utterance": "какое лекарство врач велел мне принимать утром?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-ru-advice-personal-03",
|
||
"utterance": "сколько денег я потратил на продукты в этом месяце?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-ru-advice-world-01",
|
||
"utterance": "как безопасно заменить розетку?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-ru-advice-world-02",
|
||
"utterance": "какая сейчас версия Debian stable?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-ru-advice-world-03",
|
||
"utterance": "что сегодня происходит на мировых рынках?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-personal-01",
|
||
"utterance": "which of my tasks is due before Friday?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-personal-02",
|
||
"utterance": "what dosage did my doctor tell me to take at breakfast?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-personal-03",
|
||
"utterance": "how much did I spend on groceries this month?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-world-01",
|
||
"utterance": "how should I clean a cast-iron pan?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-world-02",
|
||
"utterance": "what is the current stable release of PostgreSQL?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
{
|
||
"id": "pb-en-advice-world-03",
|
||
"utterance": "which major elections are happening this month?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "advice_current_info"
|
||
},
|
||
|
||
{
|
||
"id": "pb-ru-proper-personal-01",
|
||
"utterance": "что я записал после доклада Линуса Торвальдса?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-ru-proper-personal-02",
|
||
"utterance": "какое мнение я высказал о фильмах Куросавы?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-ru-proper-personal-03",
|
||
"utterance": "когда у меня билеты на концерт Земфиры?",
|
||
"lang": "ru",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-ru-proper-world-01",
|
||
"utterance": "кто такой Алан Тьюринг?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-ru-proper-world-02",
|
||
"utterance": "чем известна Фрида Кало?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-ru-proper-world-03",
|
||
"utterance": "когда родился Юрий Гагарин?",
|
||
"lang": "ru",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-personal-01",
|
||
"utterance": "what notes did I make after Grace Hopper's talk?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-personal-02",
|
||
"utterance": "which David Bowie album did I say I liked most?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-personal-03",
|
||
"utterance": "when are my tickets for the Radiohead show?",
|
||
"lang": "en",
|
||
"want": "personal",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-world-01",
|
||
"utterance": "who was Katherine Johnson?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-world-02",
|
||
"utterance": "what is Antoni Gaudí famous for?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
},
|
||
{
|
||
"id": "pb-en-proper-world-03",
|
||
"utterance": "when was Nelson Mandela born?",
|
||
"lang": "en",
|
||
"want": "world",
|
||
"stratum": "public_proper_nouns"
|
||
}
|
||
]
|
||
}
|