| """GPT-2 SAE last-token programs, layer 6 -- features 02400 to 02499 (100 programs). |
| |
| One function per SAE feature. Grouped 100 to a file so the set stays browsable; |
| the programs themselves are unchanged. |
| """ |
|
|
| def prog_L06F02400(input_tokens): |
| """40 token families, 55 surface forms (train F1 of the uncapped optimum: 0.685). |
| Fires at 14,177 of 36,786,566 positions.""" |
| |
| """prev SAE description: blank or non-representative sections of text""" |
| |
| fire = { |
| |
| 'og', 'OG', ' Og', ' OG', |
| |
| ' Dog', 'dog', ' Dogs', 'Dog', |
| |
| ' Adventure', 'Adventure', |
| |
| ' Oz', |
| |
| 'hog', ' hog', ' Hog', |
| |
| ' fog', ' Fog', |
| |
| 'log', 'LOG', ' LOG', |
| |
| 'ogs', |
| |
| 'oga', |
| |
| 'ogging', 'ogged', |
| |
| 'ogan', |
| |
| 'ogue', |
| |
| ' Jag', |
| |
| ' Gamer', 'Gamer', |
| |
| 'ogo', |
| |
| 'odo', |
| |
| ' Yog', ' yog', |
| |
| ' backlog', |
| |
| ' jog', |
| |
| ' bog', |
| |
| 'ggy', |
| |
| 'bug', |
| |
| 'ogle', |
| |
| ' addicts', |
| |
| ' traff', |
| |
| 'GI', |
| |
| 'gey', |
| |
| '****', |
| |
| 'ogie', |
| |
| ' Poo', |
| |
| ' slog', |
| |
| 'roach', |
| |
| ' Pig', |
| |
| ' Wag', |
| |
| 'ogl', |
| |
| 'atmeal', |
| |
| ' jug', |
| |
| 'ogi', |
| |
| ' prag', |
| |
| 'iggins', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'og-family', 1 |
|
|
| def prog_L06F02401(input_tokens): |
| """29 token families, 50 surface forms (train F1 of the uncapped optimum: 0.249). |
| Fires at 127,154 of 36,786,566 positions.""" |
| |
| """prev SAE description: descriptive adjectives that highlight positive attributes and significance""" |
| |
| fire = { |
| |
| ' new', ' newer', |
| |
| ' major', 'majority', 'major', |
| |
| ' great', ' greatest', 'great', ' Greatest', ' GREAT', |
| |
| ' big', |
| |
| ' top', ' TOP', |
| |
| ' different', |
| |
| ' important', 'important', ' Important', |
| |
| ' key', |
| |
| ' interesting', 'interesting', ' Interesting', |
| |
| ' biggest', |
| |
| ' young', |
| |
| ' favorite', 'favorite', 'Favorite', |
| |
| ' cool', ' coolest', |
| |
| ' unique', 'unique', |
| |
| ' amazing', ' amazingly', |
| |
| ' famous', ' Famous', 'famous', |
| |
| ' common', |
| |
| ' specific', 'specific', |
| |
| ' senior', |
| |
| ' awesome', |
| |
| ' significant', |
| |
| ' basic', |
| |
| ' exciting', |
| |
| ' prominent', |
| |
| ' additional', |
| |
| 'profile', |
| |
| ' notable', |
| |
| ' successful', 'successful', |
| |
| ' iconic', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'new-family', 1 |
|
|
| def prog_L06F02402(input_tokens): |
| """33 token families, 43 surface forms (train F1 of the uncapped optimum: 0.560). |
| Fires at 10,342 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to scholars and scholarly work""" |
| |
| fire = { |
| |
| ' professor', ' professors', |
| |
| ' scholars', ' scholar', ' scholarly', ' Scholar', ' Scholars', |
| |
| ' academic', |
| |
| ' historian', ' historians', |
| |
| ' philosopher', ' philosophers', |
| |
| ' academics', |
| |
| ' Studies', ' Stud', |
| |
| ' intellectual', |
| |
| ' researcher', |
| |
| ' thinkers', ' thinker', |
| |
| ' cleric', |
| |
| ' scholarship', |
| |
| ' archaeologists', |
| |
| ' academia', |
| |
| ' Brookings', |
| |
| ' intellectuals', |
| |
| ' theolog', |
| |
| ' ibn', ' Ibn', |
| |
| ' UNESCO', |
| |
| 'sch', |
| |
| 'myra', |
| |
| 'ī', |
| |
| ' clerics', |
| |
| 'disciplinary', |
| |
| ' Maul', |
| |
| 'iologist', |
| |
| 'edu', |
| |
| 'opher', |
| |
| ' disciplines', |
| |
| 'olars', |
| |
| 'ademic', |
| |
| 'orians', |
| |
| ' intellect', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'professor-family', 1 |
|
|
| def prog_L06F02403(input_tokens): |
| """1 token families, 2 surface forms (train F1 of the uncapped optimum: 0.355). |
| Fires at 137,036 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the word "and," indicating a focus on conjunctions or connections between ideas""" |
| |
| fire = { |
| |
| ' and', ' AND', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'and-family', 1 |
|
|
| def prog_L06F02404(input_tokens): |
| """39 token families, 46 surface forms (train F1 of the uncapped optimum: 0.749). |
| Fires at 8,104 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms related to racism and its manifestations""" |
| |
| fire = { |
| |
| ' racist', 'racist', ' racists', |
| |
| ' racial', ' Rac', 'rac', 'racial', ' Racial', |
| |
| ' racism', |
| |
| ' racially', |
| |
| ' supremacist', |
| |
| ' genocide', |
| |
| ' bigotry', |
| |
| ' ignorant', |
| |
| ' prejudice', |
| |
| ' discriminatory', |
| |
| ' savage', |
| |
| ' profiling', |
| |
| ' Ricky', |
| |
| 'ryan', |
| |
| ' segregation', |
| |
| ' Rodney', |
| |
| ' KKK', |
| |
| ' intolerance', |
| |
| ' ignorance', |
| |
| 'acist', 'acists', |
| |
| ' Roof', |
| |
| ' oppression', |
| |
| ' apartheid', |
| |
| ' segregated', |
| |
| ' systemic', |
| |
| 'immigration', |
| |
| ' multicultural', |
| |
| ' melan', |
| |
| ' prejudices', |
| |
| ' tolerant', |
| |
| 'igger', |
| |
| ' vile', |
| |
| ' biases', |
| |
| ' Ary', |
| |
| ' implicit', |
| |
| ' bigot', |
| |
| ' prejud', |
| |
| ' Moor', |
| |
| ' retard', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'racist-family', 1 |
|
|
| def prog_L06F02405(input_tokens): |
| """43 token families, 55 surface forms (train F1 of the uncapped optimum: 0.754). |
| Fires at 11,403 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to the Tor network or browser""" |
| |
| fire = { |
| |
| ' Tor', 'TOR', 'tor', ' tor', 'Tor', ' TOR', ' Torment', |
| |
| ' Tur', 'Tur', |
| |
| ' Tab', 'Tab', |
| |
| ' Trent', |
| |
| ' Ridge', |
| |
| ' torch', ' Torch', ' torches', |
| |
| ' Trevor', |
| |
| ' UFO', |
| |
| ' Mot', |
| |
| ' Gat', |
| |
| ' Patterson', |
| |
| ' TSA', |
| |
| ' Torres', |
| |
| 'VPN', ' VPN', |
| |
| ' Kor', |
| |
| ' Elliot', |
| |
| ' Gavin', |
| |
| ' Toro', |
| |
| ' fet', |
| |
| 'rir', |
| |
| ' Pist', ' pist', |
| |
| ' Fur', |
| |
| ' Von', |
| |
| ' Patreon', |
| |
| ' Chapel', |
| |
| ' Osh', |
| |
| ' Onion', |
| |
| ' Iz', |
| |
| ' Parsons', |
| |
| ' Juliet', |
| |
| ' torpedo', |
| |
| ' torque', |
| |
| ' Cors', |
| |
| ' PI', |
| |
| ' Trop', |
| |
| ' Tir', |
| |
| ' pier', |
| |
| ' cape', |
| |
| ' Koz', |
| |
| 'oras', |
| |
| ' DeV', |
| |
| ' PJ', |
| |
| ' Tosh', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'tor-family', 1 |
|
|
| def prog_L06F02406(input_tokens): |
| """3 token families, 4 surface forms (train F1 of the uncapped optimum: 0.863). |
| Fires at 18,378 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the letter 'L' in various contexts""" |
| |
| fire = { |
| |
| ' L', 'L', |
| |
| ' Lars', |
| |
| ' LL', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'l-family', 1 |
|
|
| def prog_L06F02407(input_tokens): |
| """50 token families, 64 surface forms (train F1 of the uncapped optimum: 0.666). |
| Fires at 33,575 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to prohibition and its implications""" |
| |
| fire = { |
| |
| ' regulation', ' Regulation', 'regulation', |
| |
| ' execution', ' Execution', |
| |
| ' destruction', ' Destruction', |
| |
| ' confirmation', |
| |
| ' exhibition', ' Exhibition', |
| |
| ' legalization', |
| |
| ' preparation', |
| |
| ' intervention', |
| |
| ' completion', |
| |
| ' repeal', |
| |
| ' prevention', ' Prevention', |
| |
| ' exploration', ' Exploration', |
| |
| ' denial', |
| |
| ' acceptance', |
| |
| ' survival', ' Survival', |
| |
| ' prohibition', ' Prohibition', |
| |
| ' refusal', |
| |
| ' revival', ' Revival', |
| |
| ' restoration', ' Restoration', |
| |
| ' isolation', |
| |
| ' medal', |
| |
| ' assassination', |
| |
| ' preservation', ' Preservation', |
| |
| ' submission', ' Submission', |
| |
| ' separation', |
| |
| ' rejection', |
| |
| ' memoir', |
| |
| ' reception', |
| |
| ' cancellation', |
| |
| ' anticipation', |
| |
| ' censorship', |
| |
| ' elimination', |
| |
| ' demolition', |
| |
| ' conversion', |
| |
| ' revision', ' Revision', |
| |
| ' obstruction', |
| |
| ' renewal', |
| |
| ' foreclosure', |
| |
| ' withdrawal', |
| |
| ' knockout', |
| |
| ' restriction', |
| |
| ' endurance', |
| |
| ' negotiation', |
| |
| ' reversal', |
| |
| 'vention', |
| |
| ' exclusion', |
| |
| ' segregation', |
| |
| ' continuation', |
| |
| ' perfection', |
| |
| ' executions', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'regula-family', 1 |
|
|
| def prog_L06F02408(input_tokens): |
| """25 token families, 31 surface forms (train F1 of the uncapped optimum: 0.228). |
| Fires at 80,569 of 36,786,566 positions.""" |
| |
| """prev SAE description: URLs and references to online content""" |
| |
| fire = { |
| |
| ' Thank', |
| |
| ' |', |
| |
| ' Show', |
| |
| ' Invalid', |
| |
| ' .', |
| |
| ' Media', |
| |
| ' Read', 'READ', |
| |
| ' https', ' http', |
| |
| ' By', |
| |
| ' +', |
| |
| ' pic', |
| |
| 'Updated', |
| |
| ' ©', '©', |
| |
| 'Posted', ' Posted', |
| |
| ' :', |
| |
| ' etc', |
| |
| 'AFP', |
| |
| ' Could', |
| |
| ' Copyright', |
| |
| ' Share', |
| |
| ' …', |
| |
| ' Stone', |
| |
| ' Video', 'VIDEO', |
| |
| ' •', |
| |
| ' plus', ' PLUS', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'thank-family', 1 |
|
|
| def prog_L06F02409(input_tokens): |
| """45 token families, 56 surface forms (train F1 of the uncapped optimum: 0.739). |
| Fires at 15,934 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to social spaces or areas where people gather, particularly lounges""" |
| |
| fire = { |
| |
| ' Boulevard', |
| |
| ' linger', ' ling', ' Ling', |
| |
| ' Lyn', 'Lyn', |
| |
| ' Lynn', |
| |
| ' drift', ' drifting', ' drifted', ' Drift', |
| |
| ' Arlington', |
| |
| ' Liberation', |
| |
| ' northwest', |
| |
| ' lounge', ' Lounge', |
| |
| ' Lor', |
| |
| ' Lloyd', |
| |
| ' Churchill', |
| |
| ' Leslie', |
| |
| ' lur', ' Lur', |
| |
| 'lake', |
| |
| ' maneuver', ' maneuvers', |
| |
| 'INGTON', |
| |
| ' laundry', |
| |
| ' Lieberman', |
| |
| ' lingering', |
| |
| ' Allison', |
| |
| ' leisure', |
| |
| ' casually', |
| |
| ' Luxembourg', |
| |
| ' remnants', ' remnant', |
| |
| ' gossip', |
| |
| ' lingu', |
| |
| ' Lexington', |
| |
| ' Bellev', |
| |
| ' LAN', |
| |
| ' reconnaissance', |
| |
| ' Redmond', |
| |
| ' lurking', |
| |
| ' Leigh', |
| |
| ' dispar', |
| |
| ' Ryzen', |
| |
| ' dwell', |
| |
| ' idle', |
| |
| ' LL', ' ll', |
| |
| ' limbo', |
| |
| ' rav', |
| |
| ' cruising', |
| |
| ' rejo', |
| |
| ' picnic', |
| |
| ' langu', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'boulevard-family', 1 |
|
|
| def prog_L06F02410(input_tokens): |
| """33 token families, 41 surface forms (train F1 of the uncapped optimum: 0.515). |
| Fires at 14,158 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms related to incarceration or imprisonment""" |
| |
| fire = { |
| |
| ' jail', ' jailed', ' jails', |
| |
| ' held', |
| |
| ' detained', ' detain', |
| |
| ' banned', |
| |
| ' imprisoned', ' imprisonment', ' imprison', |
| |
| ' hospitalized', |
| |
| ' prisoner', ' Prison', 'prison', |
| |
| ' deb', |
| |
| ' housed', |
| |
| ' exile', ' Exile', |
| |
| ' sidelined', |
| |
| ' confinement', |
| |
| ' sentences', |
| |
| ' incarceration', |
| |
| ' Sonia', |
| |
| ' outlawed', |
| |
| ' captivity', |
| |
| ' confined', |
| |
| ' incarcerated', |
| |
| ' chilling', |
| |
| ' shrouded', |
| |
| ' persecuted', |
| |
| 'rieve', |
| |
| ' hanged', |
| |
| ' paralyzed', |
| |
| ' exiled', |
| |
| ' disenfranch', |
| |
| ' captive', |
| |
| ' lecturer', |
| |
| ' implanted', |
| |
| 'alysed', |
| |
| ' enslaved', |
| |
| ' banished', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'jail-family', 1 |
|
|
| def prog_L06F02411(input_tokens): |
| """6 token families, 9 surface forms (train F1 of the uncapped optimum: 0.728). |
| Fires at 23,561 of 36,786,566 positions.""" |
| |
| """prev SAE description: phrases that indicate opposition or resistance to various issues or entities""" |
| |
| fire = { |
| |
| ' against', ' Against', 'against', |
| |
| ' denounced', ' denouncing', |
| |
| ' oppose', |
| |
| ' defeating', |
| |
| ' defund', |
| |
| ' combating', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'against-family', 1 |
|
|
| def prog_L06F02412(input_tokens): |
| """46 token families, 47 surface forms (train F1 of the uncapped optimum: 0.632). |
| Fires at 7,119 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to commonwealth entities or concepts""" |
| |
| fire = { |
| |
| ' Commonwealth', |
| |
| ' Fairfax', |
| |
| ' Carol', |
| |
| ' deceased', |
| |
| ' Caller', |
| |
| ' remainder', |
| |
| ' Caroline', |
| |
| ' Sunshine', |
| |
| ' Vanderbilt', |
| |
| 'wealth', |
| |
| ' Victorian', |
| |
| ' 365', |
| |
| 'istrates', |
| |
| ' constellation', |
| |
| ' Comm', |
| |
| ' commencement', |
| |
| ' beverage', |
| |
| ' Dominion', |
| |
| ' Mellon', |
| |
| ' Buckingham', |
| |
| ' Spectrum', |
| |
| ' Suffolk', |
| |
| ' Colleg', |
| |
| ' PEOPLE', |
| |
| ' Duchess', |
| |
| 'obile', |
| |
| ' Cooperative', |
| |
| ' Device', 'Device', |
| |
| ' Vanguard', |
| |
| ' Courier', |
| |
| 'ellation', |
| |
| ' Caucasian', |
| |
| 'allic', |
| |
| ' Pathfinder', |
| |
| ' Citadel', |
| |
| 'clave', |
| |
| ' custod', |
| |
| ' Buckley', |
| |
| 'abulary', |
| |
| ' STATES', |
| |
| ' treasury', |
| |
| ' Confederacy', |
| |
| 'shine', |
| |
| ' crimson', |
| |
| 'Austral', |
| |
| ' radiant', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'commonwealth-family', 1 |
|
|
| def prog_L06F02413(input_tokens): |
| """49 token families, 66 surface forms (train F1 of the uncapped optimum: 0.687). |
| Fires at 19,905 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to the concept of "race" or "racial issues." """ |
| |
| fire = { |
| |
| ' Vince', |
| |
| ' Trad', ' Trader', |
| |
| ' Osama', |
| |
| ' sever', 'sever', ' Sever', |
| |
| ' tid', ' Tid', |
| |
| 'tha', ' tha', |
| |
| 'ALT', |
| |
| 'oshi', |
| |
| ' USL', |
| |
| ' overhead', |
| |
| 'ando', |
| |
| ' medi', 'medi', |
| |
| ' verb', ' Verb', 'verb', ' verbs', |
| |
| ' Salv', ' salv', |
| |
| ' Lieberman', |
| |
| 'VERTIS', |
| |
| ' contr', ' CONTR', |
| |
| ' Lowe', |
| |
| 'uba', |
| |
| ' Issa', |
| |
| ' Diablo', |
| |
| 'SHARE', |
| |
| ' petty', ' Petty', |
| |
| ' ra', |
| |
| ' Premiership', |
| |
| ' temp', ' Temp', 'temp', 'Temp', |
| |
| ' Semin', ' semin', |
| |
| ' racket', |
| |
| 'Warning', |
| |
| ' McCorm', |
| |
| ' Ryzen', |
| |
| ' Truman', |
| |
| ' salvage', |
| |
| ' caveat', ' caveats', |
| |
| 'AMA', |
| |
| ' tractor', |
| |
| ' Mazda', |
| |
| ' sticky', |
| |
| ' Pulitzer', |
| |
| ' Rahman', |
| |
| ' Serious', |
| |
| ' tuna', |
| |
| ' Roku', |
| |
| ' stri', |
| |
| ' ante', |
| |
| ' Freddy', |
| |
| 'Untitled', |
| |
| ' Torch', |
| |
| ' antis', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'vince-family', 1 |
|
|
| def prog_L06F02414(input_tokens): |
| """26 token families, 51 surface forms (train F1 of the uncapped optimum: 0.308). |
| Fires at 30,150 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to researchers and authors of studies or papers""" |
| |
| fire = { |
| |
| ' author', ' authors', 'author', 'authored', 'authors', ' authored', ' AUTHOR', |
| |
| ' chairman', ' Chairman', |
| |
| ' director', 'director', |
| |
| 'founder', |
| |
| ' producer', ' Producer', |
| |
| ' Leader', 'leader', 'lead', |
| |
| 'writer', ' Writer', 'writers', |
| |
| 'chair', ' Chair', ' chaired', 'Chair', 'chairs', |
| |
| 'host', |
| |
| ' sponsor', ' sponsors', 'sponsored', |
| |
| ' researcher', |
| |
| ' architect', ' architects', |
| |
| ' contributor', ' contributors', |
| |
| ' investigator', ' Investigator', |
| |
| 'creator', |
| |
| ' speaker', |
| |
| ' collaborator', ' collaborators', |
| |
| ' coordinator', ' Coordinator', |
| |
| ' attendance', |
| |
| 'woman', |
| |
| 'seller', ' Seller', |
| |
| ' organizer', |
| |
| 'child', |
| |
| ' ingredient', |
| |
| ' scorer', |
| |
| 'owner', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'author-family', 1 |
|
|
| def prog_L06F02415(input_tokens): |
| """16 token families, 23 surface forms (train F1 of the uncapped optimum: 0.668). |
| Fires at 9,840 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the word "suffered" in relation to various contexts of distress or injuries""" |
| |
| fire = { |
| |
| ' suffered', ' suffering', |
| |
| ' suffer', ' suffers', 'suff', |
| |
| ' sustained', ' sustaining', |
| |
| ' battling', ' battled', |
| |
| ' endured', |
| |
| ' inflicted', ' inflict', ' inflicting', |
| |
| ' contracted', |
| |
| ' endure', |
| |
| ' fared', |
| |
| 'uffer', |
| |
| ' benefited', |
| |
| ' incurred', |
| |
| 'athed', |
| |
| ' crippled', |
| |
| ' fend', |
| |
| ' incur', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'suffer-family', 1 |
|
|
| def prog_L06F02416(input_tokens): |
| """13 token families, 15 surface forms (train F1 of the uncapped optimum: 0.162). |
| Fires at 22,556 of 36,786,566 positions.""" |
| |
| """prev SAE description: names and identifiers associated with individuals and groups in conflict or military contexts""" |
| |
| fire = { |
| |
| 'Assad', |
| |
| 'Click', |
| |
| 'Sh', |
| |
| 'Q', |
| |
| 'Qaeda', |
| |
| 'Nusra', |
| |
| 'Az', ' Az', |
| |
| 'Or', |
| |
| 'Sham', |
| |
| 'Qaida', |
| |
| 'Ab', |
| |
| ' Shar', 'Shar', |
| |
| 'aghd', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'assad-family', 1 |
|
|
| def prog_L06F02417(input_tokens): |
| """1 token families, 1 surface forms (train F1 of the uncapped optimum: 0.047). |
| Fires at 34,793 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to criminal charges and legal terminology""" |
| |
| fire = { |
| |
| '\n', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '\\n-family', 1 |
|
|
| def prog_L06F02418(input_tokens): |
| """48 token families, 51 surface forms (train F1 of the uncapped optimum: 0.671). |
| Fires at 10,579 of 36,786,566 positions.""" |
| |
| """prev SAE description: the occurrence of a specific character string, particularly "ette" """ |
| |
| fire = { |
| |
| 'ette', |
| |
| 'etta', |
| |
| 'ettes', 'etting', 'etter', |
| |
| 'ella', |
| |
| 'ée', |
| |
| 'ista', |
| |
| 'leton', |
| |
| 'eline', |
| |
| 'lette', |
| |
| 'anne', |
| |
| 'iley', |
| |
| 'etti', |
| |
| 'esque', |
| |
| 'otte', |
| |
| 'oko', |
| |
| 'lete', |
| |
| ' Lionel', |
| |
| 'ono', |
| |
| 'scenes', |
| |
| 'ourke', |
| |
| 'iere', |
| |
| 'onde', |
| |
| 'enne', |
| |
| 'baum', |
| |
| ' Done', 'Done', |
| |
| 'anche', |
| |
| ' palette', |
| |
| 'kered', |
| |
| 'nee', |
| |
| ' une', |
| |
| 'rosse', |
| |
| 'otom', |
| |
| 'oco', |
| |
| 'itten', |
| |
| 'enzie', |
| |
| ' cue', |
| |
| 'neau', |
| |
| 'nette', |
| |
| 'enstein', |
| |
| 'ecake', |
| |
| 'oci', |
| |
| 'itto', |
| |
| 'Ble', |
| |
| ' Babe', |
| |
| 'lique', |
| |
| ' cassette', |
| |
| 'ttes', |
| |
| 'esan', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'ette-family', 1 |
|
|
| def prog_L06F02419(input_tokens): |
| """3 token families, 5 surface forms (train F1 of the uncapped optimum: 0.397). |
| Fires at 21,729 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to the passage of time, particularly in terms of years""" |
| |
| fire = { |
| |
| ' years', 'years', ' Years', |
| |
| ' centuries', |
| |
| ' millennia', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'year-family', 1 |
|
|
| def prog_L06F02420(input_tokens): |
| """3 token families, 5 surface forms (train F1 of the uncapped optimum: 0.303). |
| Fires at 12,309 of 36,786,566 positions.""" |
| |
| """prev SAE description: the presence of the word "there" """ |
| |
| fire = { |
| |
| ' there', 'there', ' THERE', |
| |
| ' here', |
| |
| ' everywhere', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'there-family', 1 |
|
|
| def prog_L06F02421(input_tokens): |
| """23 token families, 37 surface forms (train F1 of the uncapped optimum: 0.521). |
| Fires at 55,691 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms related to roles, particularly in the context of showrunners and milestone events""" |
| |
| fire = { |
| |
| 'stone', |
| |
| 'uation', 'uations', |
| |
| 'runner', 'runners', |
| |
| ' milestone', |
| |
| 'dale', |
| |
| 'aries', 'arial', |
| |
| ' finale', |
| |
| ' candidacy', |
| |
| 'ality', |
| |
| 'estone', |
| |
| 'ailing', 'ailable', |
| |
| 'starter', 'start', 'started', 'starting', |
| |
| 'ular', 'ularity', |
| |
| ' frontrunner', |
| |
| ' Trafford', |
| |
| ' dedication', |
| |
| 'elson', |
| |
| 'ographical', 'ographies', 'ographs', |
| |
| 'graded', ' grad', 'grading', |
| |
| 'stones', ' stones', |
| |
| 'esis', 'esi', |
| |
| 'uary', |
| |
| 'ication', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'stone-family', 1 |
|
|
| def prog_L06F02422(input_tokens): |
| """38 token families, 43 surface forms (train F1 of the uncapped optimum: 0.697). |
| Fires at 6,977 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the word "unfortunate." """ |
| |
| fire = { |
| |
| ' unfortunately', ' unfortunate', |
| |
| ' tragic', |
| |
| ' fortunate', ' fortunately', 'fortunately', |
| |
| ' lamented', ' lament', |
| |
| ' disappointing', |
| |
| ' unacceptable', |
| |
| ' ironic', |
| |
| ' understandable', |
| |
| ' irresponsible', |
| |
| ' foolish', |
| |
| ' sadly', |
| |
| ' understandably', |
| |
| ' saddened', |
| |
| ' unsuspecting', |
| |
| ' egregious', |
| |
| ' regretted', ' regrett', |
| |
| ' calam', |
| |
| ' unhappy', |
| |
| ' depl', |
| |
| ' unlucky', |
| |
| ' misguided', |
| |
| ' appalling', |
| |
| ' disgruntled', |
| |
| ' exacerbated', |
| |
| ' tragically', |
| |
| ' misfortune', |
| |
| ' fateful', |
| |
| ' undesirable', |
| |
| ' shameful', |
| |
| ' ironically', |
| |
| 'ortunate', |
| |
| ' unforeseen', |
| |
| 'rehensible', |
| |
| ' perverse', |
| |
| ' despicable', |
| |
| ' troublesome', |
| |
| ' pity', |
| |
| 'uitous', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'unfortunate-family', 1 |
|
|
| def prog_L06F02423(input_tokens): |
| """10 token families, 14 surface forms (train F1 of the uncapped optimum: 0.873). |
| Fires at 21,185 of 36,786,566 positions.""" |
| |
| """prev SAE description: negative language regarding existence or certainty""" |
| |
| fire = { |
| |
| ' isn', ' Isn', 'Isn', |
| |
| ' wasn', |
| |
| ' aren', ' Aren', 'Aren', |
| |
| ' hasn', |
| |
| ' weren', |
| |
| ' hadn', |
| |
| ' shouldn', |
| |
| ' Was', |
| |
| ' OEM', |
| |
| ' huh', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'isn-family', 1 |
|
|
| def prog_L06F02424(input_tokens): |
| """16 token families, 22 surface forms (train F1 of the uncapped optimum: 0.076). |
| Fires at 386,542 of 36,786,566 positions. |
| Fires at 1.05% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: verbs and actions that indicate processes or changes""" |
| |
| fire = { |
| |
| ',', |
| |
| ' in', |
| |
| ' is', |
| |
| ' has', 'has', |
| |
| "'s", |
| |
| ' (', '(', |
| |
| ' have', 'have', |
| |
| ' will', ' willingly', |
| |
| ' or', |
| |
| ' can', |
| |
| ' —', '—', |
| |
| ' had', |
| |
| ' would', |
| |
| '?', |
| |
| ' into', |
| |
| ' –', '–', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return ',-family', 1 |
|
|
| def prog_L06F02425(input_tokens): |
| """42 token families, 48 surface forms (train F1 of the uncapped optimum: 0.733). |
| Fires at 5,240 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of specific individuals with the name "Sher" or similar variations""" |
| |
| fire = { |
| |
| ' Sher', 'Sher', ' sher', |
| |
| ' Sherman', |
| |
| 'Vict', 'vict', |
| |
| ' Siri', |
| |
| ' Sherlock', |
| |
| ' Maur', |
| |
| ' Calder', |
| |
| ' Aer', 'aer', |
| |
| ' Silent', |
| |
| ' Shel', |
| |
| ' Californ', |
| |
| 'resistant', |
| |
| ' Cul', |
| |
| ' jud', |
| |
| ' stair', |
| |
| ' Kyr', |
| |
| ' Christy', |
| |
| ' Sheridan', |
| |
| 'reci', |
| |
| 'arsh', |
| |
| ' McH', |
| |
| ' Gund', |
| |
| ' Sched', ' sched', |
| |
| ' Tul', |
| |
| ' Yak', |
| |
| ' insur', |
| |
| 'herry', |
| |
| ' Rath', |
| |
| 'sd', ' sd', |
| |
| ' undert', |
| |
| 'STE', |
| |
| ' Rhod', |
| |
| ' Tib', |
| |
| ' Saur', |
| |
| ' Hirosh', |
| |
| ' Mechan', |
| |
| 'edar', |
| |
| 'arag', |
| |
| 'sr', |
| |
| ' pilgr', |
| |
| ' sensit', |
| |
| 'cler', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'sher-family', 1 |
|
|
| def prog_L06F02426(input_tokens): |
| """11 token families, 25 surface forms (train F1 of the uncapped optimum: 0.126). |
| Fires at 364,785 of 36,786,566 positions. |
| Fires at 0.99% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: titles or names of media and publications""" |
| |
| fire = { |
| |
| ',', |
| |
| 's', "'s", ' S', 'és', 'ís', |
| |
| ':', |
| |
| ' (', |
| |
| ')', |
| |
| '/', ' /', |
| |
| ' N', 'N', 'n', 'án', 'ón', 'ín', |
| |
| ' R', 'R', 'r', ' Ré', |
| |
| "'", |
| |
| ' 3', '3', |
| |
| ' University', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return ',-family', 1 |
|
|
| def prog_L06F02427(input_tokens): |
| """27 token families, 34 surface forms (train F1 of the uncapped optimum: 0.239). |
| Fires at 342,756 of 36,786,566 positions. |
| Fires at 0.93% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: instances of prominence and emphasis related to concepts, decisions, or entities""" |
| |
| fire = { |
| |
| ' The', |
| |
| ' It', |
| |
| ' In', |
| |
| ' This', |
| |
| ' We', |
| |
| ' A', |
| |
| ' He', |
| |
| ' As', |
| |
| ' For', |
| |
| ' There', |
| |
| ' Here', 'Here', 'here', |
| |
| ' With', 'With', |
| |
| ' You', |
| |
| ' They', |
| |
| ' That', |
| |
| ' If', |
| |
| ' While', |
| |
| ' Some', 'Some', |
| |
| ' One', |
| |
| ' After', |
| |
| ' On', |
| |
| ' What', |
| |
| ' Now', |
| |
| ' When', |
| |
| ' However', |
| |
| ' First', 'First', 'Firstly', ' Firstly', |
| |
| ' So', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'the-family', 1 |
|
|
| def prog_L06F02428(input_tokens): |
| """30 token families, 36 surface forms (train F1 of the uncapped optimum: 0.664). |
| Fires at 8,447 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of a coalition in various contexts""" |
| |
| fire = { |
| |
| ' coalition', ' Coalition', |
| |
| ' alliance', |
| |
| ' consortium', ' Consortium', |
| |
| ' allied', ' Allied', ' Allies', |
| |
| ' Task', 'Task', |
| |
| ' convoy', |
| |
| ' multinational', |
| |
| ' Kristen', |
| |
| ' cartel', |
| |
| ' conglomerate', |
| |
| ' bloc', |
| |
| ' Covenant', ' covenant', |
| |
| ' alliances', |
| |
| ' Collabor', |
| |
| ' constituency', |
| |
| ' umbrella', |
| |
| ' medi', |
| |
| ' Vit', |
| |
| ' Combined', |
| |
| ' tandem', |
| |
| 'Together', |
| |
| ' constituencies', |
| |
| ' Pact', |
| |
| ' cabal', |
| |
| ' Axis', |
| |
| ' Communities', |
| |
| ' conglomer', |
| |
| 'coal', |
| |
| ' ozone', |
| |
| ' hegemony', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'coali-family', 1 |
|
|
| def prog_L06F02429(input_tokens): |
| """49 token families, 54 surface forms (train F1 of the uncapped optimum: 0.764). |
| Fires at 7,366 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the name "Tyler." """ |
| |
| fire = { |
| |
| ' Tyler', 'Tyler', |
| |
| ' Glen', |
| |
| ' Cody', |
| |
| ' fraternity', |
| |
| ' Rochester', |
| |
| ' dude', ' Dude', |
| |
| ' Tracy', |
| |
| ' Watt', ' watt', |
| |
| ' Dustin', |
| |
| ' Crosby', |
| |
| ' whiskey', |
| |
| 'essa', |
| |
| ' Connor', |
| |
| ' Eld', |
| |
| ' Kendrick', |
| |
| ' THC', |
| |
| ' Preston', |
| |
| ' Clayton', |
| |
| ' Boyle', |
| |
| 'xx', |
| |
| ' Dillon', |
| |
| ' thesis', |
| |
| ' exponentially', |
| |
| ' Valerie', |
| |
| ' Tribe', |
| |
| ' Roses', |
| |
| ' Zhang', |
| |
| ' Isaiah', |
| |
| ' cinnamon', |
| |
| ' voltage', ' Voltage', |
| |
| ' Towns', |
| |
| ' Wyatt', |
| |
| ' Poe', |
| |
| ' connector', ' connectors', |
| |
| 'Orange', |
| |
| ' Friendship', |
| |
| 'andon', |
| |
| ' condos', |
| |
| ' Tanner', |
| |
| ' Vanessa', |
| |
| ' Toll', |
| |
| ' Quincy', |
| |
| 'ofi', |
| |
| ' 1070', |
| |
| ' watered', |
| |
| 'aleb', |
| |
| 'aldehyde', |
| |
| 'ayson', |
| |
| 'ezvous', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'tyl-family', 1 |
|
|
| def prog_L06F02430(input_tokens): |
| """44 token families, 51 surface forms (train F1 of the uncapped optimum: 0.702). |
| Fires at 10,259 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the name "Morrison" and related variations, indicating a focus on a specific individual or subject""" |
| |
| fire = { |
| |
| ' Mor', ' mor', 'mor', 'Mor', ' MOR', |
| |
| ' Mormon', |
| |
| ' Morales', ' morality', ' morals', |
| |
| ' Fargo', |
| |
| ' Morrison', |
| |
| ' Morocco', |
| |
| 'Activ', |
| |
| ' mushrooms', |
| |
| ' Mayo', |
| |
| 'amine', |
| |
| ' Grateful', |
| |
| ' Moroccan', |
| |
| ' halluc', |
| |
| ' mana', ' Mana', |
| |
| ' Mori', |
| |
| ' tresp', |
| |
| ' Moran', |
| |
| ' Midnight', |
| |
| ' Nightmare', |
| |
| ' resonance', |
| |
| ' Marriott', |
| |
| ' Jenner', |
| |
| 'jer', |
| |
| ' Lowry', |
| |
| ' disturbed', |
| |
| ' psychedelic', |
| |
| 'izo', |
| |
| ' MDMA', |
| |
| ' ethanol', |
| |
| ' nighttime', |
| |
| ' Psychiatry', |
| |
| 'mare', |
| |
| ' Nietzsche', |
| |
| ' morphine', |
| |
| 'Grey', |
| |
| ' fungi', |
| |
| 'ormon', |
| |
| 'Theme', |
| |
| 'oodoo', |
| |
| ' semantic', |
| |
| 'Ore', |
| |
| ' PCI', |
| |
| ' Moonlight', |
| |
| ' Mayhem', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'mor-family', 1 |
|
|
| def prog_L06F02431(input_tokens): |
| """31 token families, 33 surface forms (train F1 of the uncapped optimum: 0.175). |
| Fires at 82,210 of 36,786,566 positions.""" |
| |
| """prev SAE description: dates and associated timestamps""" |
| |
| fire = { |
| |
| ' -', |
| |
| ' --', |
| |
| ' 2017', |
| |
| ' 2016', |
| |
| ' 2015', |
| |
| ' 2014', |
| |
| ' 12', |
| |
| ' 11', |
| |
| ' 2013', |
| |
| ' 9', |
| |
| ' /', |
| |
| ' 13', |
| |
| ' 17', |
| |
| ' 19', |
| |
| ' 14', |
| |
| ' 24', |
| |
| ' 26', |
| |
| ' 16', |
| |
| ' 23', |
| |
| ' 22', |
| |
| ' 27', |
| |
| ' GMT', |
| |
| ' 21', |
| |
| ' PM', |
| |
| ' 29', |
| |
| ' 28', |
| |
| ' 31', |
| |
| ' Feb', 'Feb', |
| |
| 'UP', |
| |
| ' Oct', ' OCT', |
| |
| ' Nov', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '--family', 1 |
|
|
| def prog_L06F02432(input_tokens): |
| """6 token families, 9 surface forms (train F1 of the uncapped optimum: 0.089). |
| Fires at 48,273 of 36,786,566 positions.""" |
| |
| """prev SAE description: individual nouns and names within various contexts""" |
| |
| fire = { |
| |
| ' A', |
| |
| ' Back', |
| |
| ' What', 'WHAT', |
| |
| ' How', ' HOW', |
| |
| ' SHARES', |
| |
| ' Your', 'Your', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'a-family', 1 |
|
|
| def prog_L06F02433(input_tokens): |
| """7 token families, 12 surface forms (train F1 of the uncapped optimum: 0.086). |
| Fires at 95,242 of 36,786,566 positions.""" |
| |
| """prev SAE description: special characters or symbols that denote various aspects of text formatting or coding""" |
| |
| fire = { |
| |
| 's', 'ís', |
| |
| ';', ' ;', |
| |
| ' -', |
| |
| ' –', '–', |
| |
| ' ...', '...', |
| |
| '…', ' …', |
| |
| 'd', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 's-family', 1 |
|
|
| def prog_L06F02434(input_tokens): |
| """7 token families, 12 surface forms (train F1 of the uncapped optimum: 0.090). |
| Fires at 298,074 of 36,786,566 positions. |
| Fires at 0.81% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: phrases and details about specific dates or timelines""" |
| |
| fire = { |
| |
| ' and', 'and', |
| |
| ' (', |
| |
| ' or', 'Or', |
| |
| ' –', '–', |
| |
| ' off', ' OFF', |
| |
| '!', |
| |
| ' under', ' Under', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'and-family', 1 |
|
|
| def prog_L06F02435(input_tokens): |
| """6 token families, 11 surface forms (train F1 of the uncapped optimum: 0.879). |
| Fires at 7,218 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to web-related topics and technologies""" |
| |
| fire = { |
| |
| ' web', ' Web', 'Web', 'web', ' webs', ' Webs', |
| |
| ' websites', |
| |
| ' WE', |
| |
| ' webcam', |
| |
| ' hamb', |
| |
| ' tv', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'web-family', 1 |
|
|
| def prog_L06F02436(input_tokens): |
| """4 token families, 5 surface forms (train F1 of the uncapped optimum: 0.079). |
| Fires at 122,277 of 36,786,566 positions.""" |
| |
| """prev SAE description: names of animal species and their classifications""" |
| |
| fire = { |
| |
| ' (', |
| |
| ' or', |
| |
| '),', ' ),', |
| |
| '—', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '(-family', 1 |
|
|
| def prog_L06F02437(input_tokens): |
| """46 token families, 51 surface forms (train F1 of the uncapped optimum: 0.608). |
| Fires at 8,778 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to molecular biology and related fields""" |
| |
| fire = { |
| |
| ' atomic', 'atomic', |
| |
| ' molecular', ' Molecular', |
| |
| ' evolutionary', |
| |
| ' Meyer', |
| |
| ' Cell', 'Cell', |
| |
| ' McMaster', |
| |
| ' laboratory', |
| |
| ' mac', |
| |
| ' orbital', ' Orbital', |
| |
| ' atmospheric', |
| |
| ' microscope', |
| |
| ' molecules', |
| |
| ' microscopic', |
| |
| ' Hubble', |
| |
| ' Chemical', |
| |
| ' cellular', |
| |
| ' macro', |
| |
| ' Pesh', |
| |
| ' computational', |
| |
| ' Baylor', |
| |
| ' muscular', |
| |
| ' Kraft', |
| |
| 'imov', |
| |
| 'neau', |
| |
| 'cular', |
| |
| ' Chemistry', |
| |
| 'anical', |
| |
| ' electron', |
| |
| ' BMC', |
| |
| ' Pearson', |
| |
| 'nec', |
| |
| ' iCloud', |
| |
| 'irie', |
| |
| ' doi', ' DOI', |
| |
| 'ochemistry', |
| |
| ' kinetic', |
| |
| ' microsc', |
| |
| 'hematic', |
| |
| ' mechan', |
| |
| ' Roche', |
| |
| ' biomedical', |
| |
| ' marrow', |
| |
| 'cule', |
| |
| ' centrif', |
| |
| ' biochemical', |
| |
| ' genomic', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'atom-family', 1 |
|
|
| def prog_L06F02438(input_tokens): |
| """48 token families, 52 surface forms (train F1 of the uncapped optimum: 0.711). |
| Fires at 18,662 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to specific sports teams, particularly "Sporting" and "Milan" """ |
| |
| fire = { |
| |
| ' Milan', |
| |
| 'Consider', ' Consider', 'consider', |
| |
| ' ankle', |
| |
| ' Drake', |
| |
| 'agan', |
| |
| 'jan', |
| |
| ' Sporting', |
| |
| 'cycl', ' Cycling', |
| |
| ' Akin', 'akin', |
| |
| 'olk', |
| |
| ' USC', |
| |
| ' Kre', |
| |
| ' Megan', |
| |
| 'edin', |
| |
| ' floated', |
| |
| 'iker', |
| |
| ' Swansea', |
| |
| 'ski', |
| |
| ' Wear', |
| |
| 'CLUS', |
| |
| ' Laz', |
| |
| ' Bryce', |
| |
| ' Timbers', |
| |
| ' Cisco', |
| |
| ' Cork', |
| |
| ' Swan', |
| |
| ' Wired', |
| |
| 'andon', |
| |
| ' Venture', |
| |
| 'avour', |
| |
| ' Cedar', |
| |
| ' Poe', |
| |
| 'dan', |
| |
| ' Stam', |
| |
| ' Keen', |
| |
| ' Erie', |
| |
| ' VMware', |
| |
| 'olen', |
| |
| ' Keane', |
| |
| 'avor', |
| |
| ' sens', |
| |
| ' opportun', |
| |
| ' Lob', |
| |
| 'oche', |
| |
| ' Ajax', |
| |
| ' Wiggins', |
| |
| 'cci', |
| |
| ' Soci', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'milan-family', 1 |
|
|
| def prog_L06F02439(input_tokens): |
| """48 token families, 51 surface forms (train F1 of the uncapped optimum: 0.721). |
| Fires at 23,090 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of specific religious and ethnic groups in the context of societal dynamics""" |
| |
| fire = { |
| |
| ' Muslims', 'Muslims', |
| |
| ' Jews', 'Jews', |
| |
| ' Christians', |
| |
| ' civilians', |
| |
| ' gamers', |
| |
| ' tourists', |
| |
| ' Palestinians', |
| |
| ' Conservatives', |
| |
| ' terrorists', |
| |
| ' minorities', |
| |
| ' Liberals', |
| |
| ' Indians', |
| |
| ' cyclists', |
| |
| ' Nazis', 'Nazis', |
| |
| ' pedestrians', |
| |
| ' Tories', |
| |
| ' Israelis', |
| |
| 'Americans', |
| |
| ' Catholics', |
| |
| ' Russians', |
| |
| ' Europeans', |
| |
| ' atheists', |
| |
| ' gays', |
| |
| ' Kurds', |
| |
| ' blacks', |
| |
| ' Marines', |
| |
| ' whites', |
| |
| ' dinosaurs', |
| |
| ' Germans', |
| |
| 'eeks', |
| |
| ' Arabs', |
| |
| ' Syrians', |
| |
| ' separatists', |
| |
| ' lesbians', |
| |
| ' Scouts', |
| |
| ' Islamists', |
| |
| ' feminists', |
| |
| ' Mexicans', |
| |
| ' Rohingya', |
| |
| ' homosexuals', |
| |
| ' Mormons', |
| |
| ' Hindus', |
| |
| ' Latinos', |
| |
| ' evangelicals', |
| |
| ' Hispanics', |
| |
| ' Armen', |
| |
| ' Asians', |
| |
| ' Iraqis', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'muslim-family', 1 |
|
|
| def prog_L06F02440(input_tokens): |
| """46 token families, 64 surface forms (train F1 of the uncapped optimum: 0.755). |
| Fires at 12,204 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of "Arm" or "Armstrong" in various contexts""" |
| |
| fire = { |
| |
| ' Arm', ' ARM', ' Arms', 'Arm', 'ARM', ' arming', |
| |
| 'Arch', ' Archer', ' Arch', |
| |
| ' Armstrong', |
| |
| ' Arnold', |
| |
| ' Emmanuel', |
| |
| ' Arlington', |
| |
| ' 82', |
| |
| ' Lance', |
| |
| ' Serv', ' servic', 'Serv', ' SERV', |
| |
| ' Batt', ' batt', |
| |
| ' Armour', ' armoured', ' armour', |
| |
| ' Barr', |
| |
| ' Atom', ' Atomic', |
| |
| ' Atari', |
| |
| ' Anglo', |
| |
| ' Aust', |
| |
| ' Bomb', |
| |
| ' Marin', |
| |
| ' Marion', |
| |
| ' stationed', |
| |
| ' Assault', 'assault', |
| |
| ' Opera', |
| |
| ' Arn', |
| |
| ' AOL', |
| |
| ' Arduino', |
| |
| ' toast', ' Toast', |
| |
| ' Ard', ' ard', |
| |
| ' Angelo', 'Angelo', |
| |
| ' Warm', |
| |
| ' Carmen', |
| |
| ' Aer', |
| |
| ' Colon', |
| |
| ' Aerospace', |
| |
| ' badass', |
| |
| ' Forge', |
| |
| ' Triumph', |
| |
| ' Advent', |
| |
| ' Hasan', |
| |
| ' Rosenberg', |
| |
| ' Germ', |
| |
| ' Blast', |
| |
| ' Sail', |
| |
| ' omnip', |
| |
| ' BAS', |
| |
| ' Aster', |
| |
| ' Ada', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'arm-family', 1 |
|
|
| def prog_L06F02441(input_tokens): |
| """40 token families, 45 surface forms (train F1 of the uncapped optimum: 0.508). |
| Fires at 10,403 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to individuals with potential or future roles, particularly related to prospects and ambitions""" |
| |
| fire = { |
| |
| ' former', |
| |
| ' potential', |
| |
| ' prospective', ' Prospect', |
| |
| ' incoming', |
| |
| ' mock', ' Mock', |
| |
| ' presumptive', |
| |
| ' aspiring', ' aspir', |
| |
| ' deceased', |
| |
| ' newborn', |
| |
| ' incumbent', |
| |
| ' applicants', ' applicant', |
| |
| ' hypothetical', |
| |
| ' unborn', |
| |
| ' budding', |
| |
| ' fian', |
| |
| ' unsuspecting', |
| |
| ' fledgling', |
| |
| ' prot', |
| |
| ' purported', |
| |
| ' predatory', |
| |
| ' retrospective', |
| |
| 'seeking', ' Seeking', |
| |
| ' fiance', |
| |
| 'ulsive', |
| |
| ' feasibility', |
| |
| ' aspire', |
| |
| ' illust', |
| |
| ' nascent', |
| |
| 'ptive', |
| |
| ' adoptive', |
| |
| ' dra', |
| |
| ' assailant', |
| |
| 'scrib', |
| |
| 'spective', |
| |
| ' groom', |
| |
| ' reincarn', |
| |
| 'Interested', |
| |
| ' bidder', |
| |
| ' Retro', |
| |
| ' observational', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'form-family', 1 |
|
|
| def prog_L06F02442(input_tokens): |
| """31 token families, 81 surface forms (train F1 of the uncapped optimum: 0.341). |
| Fires at 100,451 of 36,786,566 positions.""" |
| |
| """prev SAE description: phrases related to the establishment or creation of entities and structures""" |
| |
| fire = { |
| |
| ' up', |
| |
| ' create', ' Create', 'Create', 'create', |
| |
| ' launched', ' launch', ' launching', ' launches', |
| |
| ' open', ' opening', ' opened', 'opening', |
| |
| ' develop', ' developed', ' developing', 'develop', ' develops', |
| |
| ' created', ' creating', ' Creating', |
| |
| ' establish', ' established', ' establishing', ' establishment', ' establishes', 'establish', |
| |
| ' hold', ' holding', |
| |
| ' conduct', ' conducted', ' conducting', ' conducts', ' Conduct', 'conduct', |
| |
| ' form', ' formed', ' forming', |
| |
| ' build', |
| |
| ' host', ' hosting', ' hosts', 'host', |
| |
| ' write', ' Write', 'write', 'Write', |
| |
| ' planning', |
| |
| ' filed', ' filing', |
| |
| ' together', |
| |
| ' agreed', |
| |
| ' founded', ' founding', 'founded', |
| |
| ' begin', 'begin', |
| |
| ' produce', |
| |
| ' recorded', ' recording', |
| |
| ' preparing', ' prepares', ' Prepar', |
| |
| ' creation', |
| |
| ' introduce', |
| |
| ' construct', ' constructing', ' constructed', 'construct', 'Construct', |
| |
| ' compiled', ' compiling', |
| |
| ' lay', ' laying', |
| |
| ' begun', |
| |
| ' staged', ' staging', |
| |
| ' commit', ' commits', |
| |
| ' producing', ' produces', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'up-family', 1 |
|
|
| def prog_L06F02443(input_tokens): |
| """2 token families, 8 surface forms (train F1 of the uncapped optimum: 0.316). |
| Fires at 29,114 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the word "have" in various contexts""" |
| |
| fire = { |
| |
| ' have', ' Have', 'have', 'Have', ' HAVE', |
| |
| ' having', ' Having', 'having', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'have-family', 1 |
|
|
| def prog_L06F02444(input_tokens): |
| """20 token families, 30 surface forms (train F1 of the uncapped optimum: 0.642). |
| Fires at 12,228 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of drugs and their related effects or approvals""" |
| |
| fire = { |
| |
| ' drug', ' drugs', ' Drug', 'drug', ' Drugs', 'Drug', |
| |
| ' vaccine', |
| |
| ' pharmaceutical', ' Pharmaceutical', |
| |
| ' pill', |
| |
| ' treatments', |
| |
| ' medicines', |
| |
| 'killer', 'killers', |
| |
| ' medications', |
| |
| ' derivative', ' derivatives', |
| |
| ' antibiotic', |
| |
| ' therapies', |
| |
| ' antidepressant', ' antidepressants', |
| |
| ' molecule', |
| |
| ' therapeutic', |
| |
| ' inhibitors', ' inhibitor', |
| |
| ' receptor', |
| |
| ' pharmac', |
| |
| 'otom', |
| |
| ' contraceptive', |
| |
| 'rugged', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'drug-family', 1 |
|
|
| def prog_L06F02445(input_tokens): |
| """18 token families, 31 surface forms (train F1 of the uncapped optimum: 0.173). |
| Fires at 218,352 of 36,786,566 positions. |
| Fires at 0.59% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: terms related to business and organizational roles""" |
| |
| fire = { |
| |
| ' system', ' systems', ' System', |
| |
| ' team', |
| |
| ' program', ' Program', |
| |
| ' series', ' Series', |
| |
| ' campaign', ' campaigns', |
| |
| ' notice', |
| |
| ' home', |
| |
| ' page', |
| |
| ' account', ' accounts', ' Account', |
| |
| ' service', |
| |
| ' policy', ' Policy', |
| |
| ' business', |
| |
| ' career', ' careers', |
| |
| ' plan', ' Plan', |
| |
| ' address', ' Address', |
| |
| ' debut', |
| |
| ' network', ' networks', |
| |
| ' committee', ' Committee', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'system-family', 1 |
|
|
| def prog_L06F02446(input_tokens): |
| """35 token families, 39 surface forms (train F1 of the uncapped optimum: 0.726). |
| Fires at 9,169 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms associated with lethal threats or serious harm""" |
| |
| fire = { |
| |
| ' deadly', ' Deadly', |
| |
| ' fatal', ' Fatal', |
| |
| ' devastating', |
| |
| ' delicious', |
| |
| ' lethal', 'lethal', ' Lethal', |
| |
| ' hilarious', |
| |
| 'threatening', |
| |
| ' deadliest', |
| |
| ' volatile', |
| |
| ' fatally', |
| |
| ' tasty', |
| |
| 'lived', |
| |
| ' formidable', |
| |
| ' unpredictable', |
| |
| ' nasty', |
| |
| ' murderous', |
| |
| ' debilitating', |
| |
| ' potent', |
| |
| ' ruthless', |
| |
| ' unstoppable', |
| |
| ' carb', |
| |
| ' Dangerous', |
| |
| ' hateful', |
| |
| ' delightful', |
| |
| ' fearsome', |
| |
| ' Firearms', |
| |
| ' lifes', |
| |
| 'killing', |
| |
| ' dazzling', |
| |
| ' irresistible', |
| |
| 'lasting', |
| |
| ' cowardly', |
| |
| ' poisonous', |
| |
| ' dodged', |
| |
| 'ESH', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'dead-family', 1 |
|
|
| def prog_L06F02447(input_tokens): |
| """12 token families, 29 surface forms (train F1 of the uncapped optimum: 0.141). |
| Fires at 125,006 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to specific characters or names in a fictional or narrative context""" |
| |
| fire = { |
| |
| '-', |
| |
| ' One', 'ONE', ' Ones', |
| |
| 'z', ' Z', 'Z', |
| |
| 'man', ' Man', 'MAN', 'mans', |
| |
| 'on', 'ON', |
| |
| '11', |
| |
| ' County', ' COUNTY', |
| |
| 'land', 'LAND', 'lander', |
| |
| ' State', ' STATE', |
| |
| ' S', |
| |
| 'ray', 'rays', 'Ray', ' rays', ' ray', |
| |
| 'ar', ' Ar', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '--family', 1 |
|
|
| def prog_L06F02448(input_tokens): |
| """46 token families, 55 surface forms (train F1 of the uncapped optimum: 0.732). |
| Fires at 15,284 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the substring "abi" related to individuals' names or references""" |
| |
| fire = { |
| |
| 'abi', 'abies', 'abis', |
| |
| ' Brisbane', |
| |
| 'abs', |
| |
| 'ifa', |
| |
| 'iki', |
| |
| ' bush', ' bushes', 'bush', |
| |
| 'aby', |
| |
| 'illo', |
| |
| ' Pak', 'Pak', |
| |
| ' Baghdad', |
| |
| ' Frankfurt', |
| |
| ' Shi', |
| |
| 'fa', |
| |
| 'iba', |
| |
| 'asi', |
| |
| 'Kit', |
| |
| ' frag', ' Frag', 'Frag', |
| |
| 'ihad', |
| |
| ' PAX', ' Pax', |
| |
| ' Carly', |
| |
| ' pillar', ' pillars', |
| |
| 'abin', |
| |
| ' McCabe', |
| |
| 'abl', |
| |
| 'ammu', |
| |
| 'heimer', |
| |
| ' Bom', |
| |
| 'ibi', |
| |
| 'auga', |
| |
| ' McKenzie', |
| |
| 'afi', |
| |
| ' Fig', |
| |
| ' Zeit', |
| |
| ' affordability', |
| |
| ' barric', |
| |
| 'pei', |
| |
| 'ibu', |
| |
| 'abo', |
| |
| 'qa', |
| |
| ' apolog', |
| |
| ' Buenos', |
| |
| 'apo', |
| |
| ' hijab', |
| |
| ' Mum', |
| |
| ' Garage', |
| |
| ' Peg', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'abi-family', 1 |
|
|
| def prog_L06F02449(input_tokens): |
| """20 token families, 26 surface forms (train F1 of the uncapped optimum: 0.775). |
| Fires at 6,045 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to tangible or physical materials""" |
| |
| fire = { |
| |
| ' material', ' Material', 'material', 'Material', |
| |
| ' materials', ' Materials', ' materially', |
| |
| ' pornography', |
| |
| ' debris', |
| |
| ' Khalid', |
| |
| ' manuscript', |
| |
| ' trivial', |
| |
| ' Hed', |
| |
| ' plagiar', |
| |
| 'IAL', |
| |
| ' collateral', |
| |
| ' fodder', |
| |
| ' rubbish', |
| |
| 'stantial', |
| |
| ' 370', |
| |
| 'aterial', 'ateral', |
| |
| '----------------------------------------------------------------', |
| |
| 'uscript', |
| |
| 'bish', |
| |
| ' nihil', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'mater-family', 1 |
|
|
| def prog_L06F02450(input_tokens): |
| """25 token families, 41 surface forms (train F1 of the uncapped optimum: 0.205). |
| Fires at 37,420 of 36,786,566 positions.""" |
| |
| """prev SAE description: phrases that involve references to specific topics or subjects""" |
| |
| fire = { |
| |
| ' about', |
| |
| ' citing', ' cited', ' cites', |
| |
| ' referring', ' referred', |
| |
| ' speaking', |
| |
| ' saying', ' Saying', |
| |
| ' adding', |
| |
| ' reference', |
| |
| ' describing', |
| |
| ' calling', 'Calling', ' Calling', |
| |
| ' noting', |
| |
| ' pointing', ' pointed', |
| |
| ' suggesting', |
| |
| ' discussing', |
| |
| 'iting', |
| |
| ' meaning', 'meaning', |
| |
| ' quoting', ' quoted', |
| |
| ' marking', |
| |
| ' accusing', |
| |
| ' explaining', |
| |
| ' refers', ' refer', 'Refer', ' Refer', |
| |
| ' referencing', ' referenced', |
| |
| ' recalling', ' recalled', ' recalls', |
| |
| ' comparing', ' compares', |
| |
| ' indicating', |
| |
| ' representing', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'about-family', 1 |
|
|
| def prog_L06F02451(input_tokens): |
| """29 token families, 50 surface forms (train F1 of the uncapped optimum: 0.297). |
| Fires at 153,576 of 36,786,566 positions. |
| Fires at 0.42% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: proper nouns, particularly names and places""" |
| |
| fire = { |
| |
| ' Al', ' al', 'Al', |
| |
| ' N', |
| |
| ' S', |
| |
| ' K', |
| |
| ' R', |
| |
| ' Sh', 'Sh', ' sh', |
| |
| ' B', |
| |
| ' P', |
| |
| ' H', |
| |
| ' M', ' Mé', ' Mü', |
| |
| ' J', |
| |
| ' T', |
| |
| ' G', |
| |
| ' Y', |
| |
| ' El', 'El', ' EL', |
| |
| ' Ch', |
| |
| ' Z', |
| |
| ' An', ' AN', |
| |
| ' Mar', 'Mar', |
| |
| 'Ad', ' Ad', |
| |
| ' Kim', 'Kim', |
| |
| ' Ab', 'Ab', |
| |
| ' Q', ' q', |
| |
| ' Ar', |
| |
| 'bel', ' Bel', 'Bel', |
| |
| ' Gu', 'Gu', ' GU', |
| |
| ' Le', |
| |
| ' Su', 'Su', ' SU', |
| |
| ' Am', 'Am', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'al-family', 1 |
|
|
| def prog_L06F02452(input_tokens): |
| """30 token families, 44 surface forms (train F1 of the uncapped optimum: 0.650). |
| Fires at 10,624 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to soldiers""" |
| |
| fire = { |
| |
| ' soldiers', ' soldier', ' Soldier', ' Soldiers', |
| |
| ' troops', |
| |
| 'sold', 'Sold', ' SOLD', ' Sold', |
| |
| ' gunfire', |
| |
| ' infantry', ' Infantry', |
| |
| ' servic', ' Serv', 'Serv', ' SERV', |
| |
| ' sergeant', ' Sergeant', |
| |
| ' sailors', ' sailor', |
| |
| ' Colonel', ' colonel', |
| |
| ' policeman', |
| |
| ' pilots', |
| |
| ' Lieutenant', |
| |
| ' settlers', |
| |
| ' warriors', |
| |
| ' trenches', ' trench', |
| |
| ' policemen', |
| |
| 'emen', |
| |
| ' Sniper', |
| |
| 'opers', |
| |
| ' generals', |
| |
| ' Spartans', |
| |
| ' comrades', |
| |
| ' medic', |
| |
| ' laborers', |
| |
| ' Riders', |
| |
| ' straps', |
| |
| ' patriotism', |
| |
| ' knight', |
| |
| 'rers', |
| |
| 'onel', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'soldi-family', 1 |
|
|
| def prog_L06F02453(input_tokens): |
| """48 token families, 66 surface forms (train F1 of the uncapped optimum: 0.665). |
| Fires at 13,739 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to computer hardware components, particularly related to sockets""" |
| |
| fire = { |
| |
| ' Pier', ' pier', |
| |
| ' Podesta', |
| |
| ' Stockholm', |
| |
| ' lodged', |
| |
| ' Stefan', |
| |
| ' Thread', ' threaded', 'Thread', 'thread', |
| |
| ' socks', ' sock', |
| |
| 'ocket', 'ockets', |
| |
| ' socket', ' sockets', 'Socket', 'socket', ' Socket', |
| |
| ' Tayyip', |
| |
| ' Kremlin', |
| |
| 'OCK', |
| |
| ' Jenny', |
| |
| ' stud', ' Stud', |
| |
| ' Tun', 'tun', |
| |
| ' Ear', |
| |
| ' Junction', ' junction', |
| |
| 'zza', |
| |
| ' sensory', |
| |
| ' knot', ' Knot', |
| |
| ' Slater', |
| |
| ' Parsons', |
| |
| 'ija', |
| |
| ' upt', |
| |
| ' Latvia', |
| |
| ' nick', |
| |
| ' restart', |
| |
| ' Loch', |
| |
| ' Peaks', |
| |
| ' Ramsey', |
| |
| ' Cer', |
| |
| ' reflex', ' Reflex', |
| |
| 'kt', |
| |
| ' TCP', ' tcp', |
| |
| 'č', |
| |
| ' Iz', |
| |
| ' cert', |
| |
| ' saint', |
| |
| ' Jenna', |
| |
| ' piercing', ' Piercing', |
| |
| ' SIG', ' sig', |
| |
| ' Esk', |
| |
| ' Ved', |
| |
| ' Eston', |
| |
| ' torpedo', |
| |
| ' Kund', |
| |
| ' Timber', |
| |
| ' serotonin', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'pier-family', 1 |
|
|
| def prog_L06F02454(input_tokens): |
| """43 token families, 53 surface forms (train F1 of the uncapped optimum: 0.614). |
| Fires at 6,928 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to fonts and typography""" |
| |
| fire = { |
| |
| ' Font', ' font', ' fonts', 'font', 'Font', |
| |
| ' Sans', ' sans', |
| |
| ' Hebdo', |
| |
| ' Pant', |
| |
| ' typ', |
| |
| ' pasta', |
| |
| ' alphabet', |
| |
| ' icons', 'icons', |
| |
| 'hens', |
| |
| ' hosp', ' Hosp', |
| |
| ' Papa', |
| |
| 'char', ' chars', |
| |
| ' hem', |
| |
| ' Gustav', |
| |
| ' encoding', |
| |
| ' handwriting', |
| |
| 'Except', |
| |
| ' SSL', |
| |
| 'BG', ' BG', 'bg', |
| |
| 'ricting', |
| |
| ' thickness', |
| |
| ' bras', |
| |
| ' Unicode', |
| |
| ' icing', |
| |
| ' gown', |
| |
| ' Gothic', |
| |
| ' gast', |
| |
| 'strap', |
| |
| ' codec', |
| |
| ' Bib', |
| |
| 'ibli', |
| |
| ' brim', |
| |
| 'igraph', |
| |
| 'faces', |
| |
| ' Franç', |
| |
| 'ONT', |
| |
| ' Hatt', |
| |
| ' pron', |
| |
| 'ilingual', |
| |
| ' glyph', |
| |
| 'ospels', |
| |
| 'ospital', |
| |
| ' conson', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'font-family', 1 |
|
|
| def prog_L06F02455(input_tokens): |
| """40 token families, 53 surface forms (train F1 of the uncapped optimum: 0.809). |
| Fires at 5,125 of 36,786,566 positions.""" |
| |
| """prev SAE description: programming syntax elements related to functions and error handling""" |
| |
| fire = { |
| |
| ' ..........', |
| |
| '")', |
| |
| ');', ' );', |
| |
| '])', ' ])', |
| |
| '"?', |
| |
| '"),', |
| |
| ')?', |
| |
| '];', ' ];', |
| |
| '))', ' ))', |
| |
| '";', |
| |
| '").', |
| |
| '"]', |
| |
| '">', |
| |
| '}}', ' }}', |
| |
| '/)', |
| |
| "')", |
| |
| ' ();', '();', |
| |
| ']),', |
| |
| ' />', '/>', |
| |
| '"></', |
| |
| ']]', |
| |
| ')]', |
| |
| ' };', '};', |
| |
| ' });', '});', |
| |
| '.")', |
| |
| '*)', ' *)', |
| |
| ' [];', |
| |
| "';", |
| |
| ' */', '*/', |
| |
| "'?", |
| |
| '/?', |
| |
| ')))', ' )))', |
| |
| '%;', |
| |
| '。', |
| |
| '}}}', |
| |
| ' @@', '@@', |
| |
| '(),', |
| |
| ')).', |
| |
| ' &&', |
| |
| ' ],', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '..........-family', 1 |
|
|
| def prog_L06F02456(input_tokens): |
| """24 token families, 34 surface forms (train F1 of the uncapped optimum: 0.247). |
| Fires at 38,160 of 36,786,566 positions.""" |
| |
| """prev SAE description: phrases describing geographical locations and their specific parts or edges""" |
| |
| fire = { |
| |
| ' side', ' Side', |
| |
| ' end', |
| |
| ' area', |
| |
| ' coast', ' Coast', ' coasts', |
| |
| ' floor', ' floors', 'floor', ' Floor', |
| |
| ' corner', ' corners', |
| |
| ' region', |
| |
| ' border', |
| |
| ' edge', |
| |
| ' section', |
| |
| ' quarter', |
| |
| ' district', |
| |
| ' phase', |
| |
| ' edition', |
| |
| ' door', |
| |
| ' outskirts', |
| |
| ' neighborhood', |
| |
| ' portion', |
| |
| ' stages', |
| |
| ' tier', ' tiers', 'tier', |
| |
| 'lines', |
| |
| 'bound', |
| |
| ' category', 'category', |
| |
| 'most', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'side-family', 1 |
|
|
| def prog_L06F02457(input_tokens): |
| """8 token families, 19 surface forms (train F1 of the uncapped optimum: 0.066). |
| Fires at 121,257 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of yelling or screaming in various contexts""" |
| |
| fire = { |
| |
| ' at', |
| |
| ' out', 'out', ' Out', 'outing', ' OUT', ' outs', 'outed', |
| |
| ' about', |
| |
| ' during', 'during', |
| |
| ' over', ' OVER', |
| |
| ' him', |
| |
| ' under', 'unders', |
| |
| ' same', ' Same', 'Same', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'at-family', 1 |
|
|
| def prog_L06F02458(input_tokens): |
| """1 token families, 1 surface forms (train F1 of the uncapped optimum: 0.536). |
| Fires at 272,759 of 36,786,566 positions. |
| Fires at 0.74% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: the article "a" in various contexts""" |
| |
| fire = { |
| |
| ' a', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'a-family', 1 |
|
|
| def prog_L06F02459(input_tokens): |
| """47 token families, 56 surface forms (train F1 of the uncapped optimum: 0.628). |
| Fires at 12,707 of 36,786,566 positions.""" |
| |
| """prev SAE description: expressions of fear and anxiety""" |
| |
| fire = { |
| |
| 'read', 'READ', |
| |
| ' dread', ' dreaded', ' Dread', ' dreadful', |
| |
| ' anticipation', |
| |
| ' Domin', 'Domin', ' domin', |
| |
| ' Von', |
| |
| ' Tyr', ' tyr', |
| |
| 'fare', |
| |
| ' Hilton', |
| |
| ' tentative', |
| |
| 'onder', |
| |
| ' Borg', |
| |
| ' vaguely', |
| |
| ' bliss', ' Bliss', |
| |
| ' Preston', |
| |
| 'athom', |
| |
| ' Hann', |
| |
| 'rence', |
| |
| ' Darth', |
| |
| ' tun', ' Tun', |
| |
| ' painfully', |
| |
| ' mull', |
| |
| ' heed', |
| |
| 'isson', |
| |
| ' countdown', |
| |
| ' agon', |
| |
| ' tread', |
| |
| ' indec', |
| |
| ' horrors', |
| |
| 'omsday', |
| |
| ' Warhammer', |
| |
| ' excruciating', |
| |
| ' suspense', |
| |
| ' Tinder', |
| |
| ' priv', |
| |
| ' McKenna', |
| |
| ' lament', |
| |
| ' Kathryn', |
| |
| ' sweats', |
| |
| ' Wyn', |
| |
| ' unaccount', |
| |
| ' WOR', |
| |
| ' torment', |
| |
| ' Cullen', |
| |
| 'dain', |
| |
| ' rotting', |
| |
| ' sleeper', |
| |
| ' Viz', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'read-family', 1 |
|
|
| def prog_L06F02460(input_tokens): |
| """1 token families, 3 surface forms (train F1 of the uncapped optimum: 0.794). |
| Fires at 107,742 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the word "at" followed by numerical values indicating times""" |
| |
| fire = { |
| |
| ' at', 'At', ' At', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'at-family', 1 |
|
|
| def prog_L06F02461(input_tokens): |
| """39 token families, 59 surface forms (train F1 of the uncapped optimum: 0.311). |
| Fires at 94,837 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to fundamental concepts or principles in various contexts""" |
| |
| fire = { |
| |
| ' results', |
| |
| ' details', |
| |
| ' issues', 'issues', |
| |
| ' rules', ' Rules', 'rules', |
| |
| ' effects', 'effects', |
| |
| ' words', |
| |
| ' numbers', |
| |
| ' events', |
| |
| ' questions', |
| |
| ' benefits', ' Benefits', |
| |
| ' reasons', ' reasoning', |
| |
| ' problems', |
| |
| ' ones', |
| |
| ' names', |
| |
| ' elements', ' Elements', |
| |
| ' challenges', ' Challenges', |
| |
| ' facts', ' Facts', 'facts', |
| |
| ' lines', 'lines', |
| |
| ' findings', |
| |
| ' principles', ' Principles', |
| |
| ' ideas', |
| |
| ' steps', ' Steps', |
| |
| ' costs', ' Costs', |
| |
| ' aspects', |
| |
| ' tools', 'tools', |
| |
| ' conditions', |
| |
| ' consequences', |
| |
| ' values', ' Values', 'values', |
| |
| ' decisions', |
| |
| ' deaths', |
| |
| ' skills', ' Skills', |
| |
| ' risks', |
| |
| ' differences', |
| |
| ' techniques', ' Techniques', |
| |
| ' factors', ' Factors', |
| |
| ' dangers', |
| |
| ' answers', |
| |
| ' requirements', |
| |
| ' opinions', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'result-family', 1 |
|
|
| def prog_L06F02462(input_tokens): |
| """5 token families, 6 surface forms (train F1 of the uncapped optimum: 0.041). |
| Fires at 18,421 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of the phrase "Back in the" followed by a year or period""" |
| |
| fire = { |
| |
| ' in', ' IN', |
| |
| ' until', |
| |
| ' December', |
| |
| ' February', |
| |
| ' 2007', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'in-family', 1 |
|
|
| def prog_L06F02463(input_tokens): |
| """19 token families, 53 surface forms (train F1 of the uncapped optimum: 0.170). |
| Fires at 678,083 of 36,786,566 positions. |
| Fires at 1.84% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: violent incidents involving injury or death""" |
| |
| fire = { |
| |
| ' a', |
| |
| ' in', |
| |
| ' on', |
| |
| ' after', |
| |
| ' at', |
| |
| ' his', |
| |
| ' from', |
| |
| ' who', |
| |
| ' her', 'hered', |
| |
| ' man', ' mansion', ' Mansion', ' mans', |
| |
| 'old', 'Old', 'olded', |
| |
| ' during', 'during', ' dur', |
| |
| ' killed', ' killing', ' killer', ' Killed', ' killings', ' kills', ' killers', ' Kills', ' Killing', 'killed', |
| |
| ' police', 'police', |
| |
| ' last', |
| |
| ' near', ' nearly', ' Nearly', ' Near', |
| |
| ' former', ' Former', 'Former', |
| |
| ' allegedly', ' alleged', ' alleging', ' Alleg', |
| |
| ' attack', ' attacked', ' attacks', ' attacking', ' attacker', ' attackers', ' Attack', 'attack', ' Attacks', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'a-family', 1 |
|
|
| def prog_L06F02464(input_tokens): |
| """19 token families, 28 surface forms (train F1 of the uncapped optimum: 0.144). |
| Fires at 335,239 of 36,786,566 positions. |
| Fires at 0.91% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: phrases that indicate necessity, capability, or ongoing actions""" |
| |
| fire = { |
| |
| ' is', |
| |
| ' has', |
| |
| ' have', |
| |
| ' will', |
| |
| ' are', |
| |
| ' had', 'had', |
| |
| ' can', |
| |
| ' would', 'would', |
| |
| ' seems', ' seem', ' seemed', |
| |
| ' may', |
| |
| 'll', "'ll", |
| |
| ' found', |
| |
| 're', |
| |
| ' still', |
| |
| ' need', |
| |
| ' took', |
| |
| ' appears', ' appeared', ' appear', |
| |
| ' find', ' finds', |
| |
| ' shows', ' showed', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'is-family', 1 |
|
|
| def prog_L06F02465(input_tokens): |
| """1 token families, 1 surface forms (train F1 of the uncapped optimum: 0.162). |
| Fires at 125,362 of 36,786,566 positions.""" |
| |
| """prev SAE description: numerical values or statistics related to player contracts and performances in sports contexts""" |
| |
| fire = { |
| |
| '\n', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '\\n-family', 1 |
|
|
| def prog_L06F02466(input_tokens): |
| """7 token families, 10 surface forms (train F1 of the uncapped optimum: 0.036). |
| Fires at 54,259 of 36,786,566 positions.""" |
| |
| """prev SAE description: instances of waking up or discovering something upon waking""" |
| |
| fire = { |
| |
| ' with', |
| |
| ' into', 'into', |
| |
| ' after', 'after', |
| |
| ' your', ' Your', |
| |
| ' find', |
| |
| 'ly', |
| |
| ' without', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'with-family', 1 |
|
|
| def prog_L06F02467(input_tokens): |
| """19 token families, 19 surface forms (train F1 of the uncapped optimum: 0.661). |
| Fires at 3,091 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to Brussels""" |
| |
| fire = { |
| |
| ' Brussels', |
| |
| 'ELS', |
| |
| ' Lisbon', |
| |
| ' fren', |
| |
| ' Lansing', |
| |
| ' Bac', |
| |
| 'sels', |
| |
| ' Stras', |
| |
| 'bourg', |
| |
| ' WTO', |
| |
| ' Unic', |
| |
| 'aucus', |
| |
| ' broccoli', |
| |
| ' DG', |
| |
| ' Brus', |
| |
| 'hent', |
| |
| ' Brill', |
| |
| 'ccoli', |
| |
| 'ingen', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'brussel-family', 1 |
|
|
| def prog_L06F02468(input_tokens): |
| """12 token families, 15 surface forms (train F1 of the uncapped optimum: 0.106). |
| Fires at 232,946 of 36,786,566 positions. |
| Fires at 0.63% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: prepositions and conjunctions that indicate relationships and conditions""" |
| |
| fire = { |
| |
| ' in', |
| |
| ' by', |
| |
| ' over', |
| |
| ' than', |
| |
| ' today', |
| |
| ' among', 'among', |
| |
| ' until', 'until', |
| |
| '—', |
| |
| ' us', 'US', |
| |
| ' Wednesday', |
| |
| ' however', |
| |
| 'g', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'in-family', 1 |
|
|
| def prog_L06F02469(input_tokens): |
| """41 token families, 48 surface forms (train F1 of the uncapped optimum: 0.589). |
| Fires at 7,909 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to specific words or phrases in a language, potentially indicating a linguistic focus or context""" |
| |
| fire = { |
| |
| 'tor', 'TOR', ' TOR', 'Tor', |
| |
| ' Pier', ' pier', |
| |
| ' Pierce', |
| |
| ' Torres', |
| |
| 'к', |
| |
| ' Kap', |
| |
| 'ski', |
| |
| 'uv', 'UV', |
| |
| 'IRO', |
| |
| ' Ivan', |
| |
| 'aja', |
| |
| 'ن', |
| |
| 'ava', |
| |
| 'izon', |
| |
| 'irus', |
| |
| 'м', |
| |
| ' UTC', |
| |
| ' punishable', |
| |
| 'в', |
| |
| 'gor', |
| |
| ' Siege', |
| |
| ' Warwick', |
| |
| 'zar', |
| |
| 'Screenshot', |
| |
| ',-', |
| |
| 'ン', |
| |
| ' Serbian', |
| |
| 'itar', |
| |
| 'ierce', |
| |
| 'px', |
| |
| 'ع', |
| |
| 'س', |
| |
| 'ル', |
| |
| 'ariat', |
| |
| 'ー�', |
| |
| 'Flag', 'flag', 'flags', |
| |
| 'י�', |
| |
| 'し', |
| |
| 'ר', |
| |
| '大', |
| |
| 'iege', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'tor-family', 1 |
|
|
| def prog_L06F02470(input_tokens): |
| """25 token families, 29 surface forms (train F1 of the uncapped optimum: 0.494). |
| Fires at 5,340 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to low-income individuals and families""" |
| |
| fire = { |
| |
| ' income', 'income', |
| |
| ' poor', ' poorest', ' poorer', |
| |
| 'quality', |
| |
| ' wealthiest', ' wealthier', |
| |
| 'ICH', |
| |
| ' impoverished', |
| |
| 'bedroom', |
| |
| 'torn', |
| |
| ' affluent', |
| |
| 'wage', |
| |
| ' earners', |
| |
| ' disadvantaged', |
| |
| ' socio', |
| |
| 'prime', |
| |
| 'economic', |
| |
| ' incomes', |
| |
| ' socioeconomic', |
| |
| 'bodied', |
| |
| 'overty', |
| |
| 'colored', |
| |
| 'ileged', |
| |
| 'represented', |
| |
| 'ticket', |
| |
| 'skilled', |
| |
| 'skinned', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'income-family', 1 |
|
|
| def prog_L06F02471(input_tokens): |
| """46 token families, 63 surface forms (train F1 of the uncapped optimum: 0.798). |
| Fires at 63,367 of 36,786,566 positions.""" |
| |
| """prev SAE description: names of U.S. states""" |
| |
| fire = { |
| |
| ' Texas', 'Texas', |
| |
| ' California', 'California', |
| |
| ' York', |
| |
| ' Florida', 'Florida', |
| |
| ' Ohio', 'Ohio', |
| |
| ' Virginia', 'Virginia', |
| |
| ' Michigan', 'Michigan', |
| |
| ' Carolina', |
| |
| ' Arizona', 'Arizona', |
| |
| ' Colorado', 'Colorado', |
| |
| ' Minnesota', 'Minnesota', |
| |
| ' Oregon', 'Oregon', |
| |
| ' Alabama', 'Alabama', |
| |
| ' Jersey', |
| |
| ' Georgia', 'Georgia', |
| |
| ' Iowa', 'Iowa', |
| |
| ' Utah', 'Utah', |
| |
| ' Wisconsin', 'Wisconsin', |
| |
| ' Kansas', 'Kansas', |
| |
| ' Illinois', |
| |
| ' Pennsylvania', |
| |
| ' Indiana', 'Indiana', |
| |
| ' Maryland', |
| |
| ' Massachusetts', |
| |
| ' Missouri', |
| |
| ' Tennessee', |
| |
| ' Kentucky', |
| |
| ' Nevada', |
| |
| ' Oklahoma', |
| |
| ' Louisiana', |
| |
| ' Alaska', |
| |
| ' Vermont', |
| |
| ' Connecticut', |
| |
| ' Arkansas', |
| |
| ' Maine', |
| |
| ' Ontario', |
| |
| ' Hawaii', |
| |
| ' Mississippi', |
| |
| ' Dakota', |
| |
| ' Hampshire', |
| |
| 'Washington', |
| |
| ' Montana', |
| |
| ' Delaware', |
| |
| ' Nebraska', |
| |
| ' Idaho', |
| |
| ' Wyoming', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'texa-family', 1 |
|
|
| def prog_L06F02472(input_tokens): |
| """23 token families, 55 surface forms (train F1 of the uncapped optimum: 0.063). |
| Fires at 420,752 of 36,786,566 positions. |
| Fires at 1.14% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: specific assertions or definitive statements regarding conditions or situations""" |
| |
| fire = { |
| |
| ' the', 'The', |
| |
| ' is', |
| |
| ' it', ' It', 'It', 'IT', |
| |
| ' has', |
| |
| ' was', |
| |
| ' I', 'I', |
| |
| ' have', ' Have', 'Have', 'have', |
| |
| ' you', ' You', 'You', |
| |
| ' are', ' Are', |
| |
| ' his', ' His', |
| |
| ' he', |
| |
| ' said', 'said', ' Said', |
| |
| ' not', 'Notable', 'NOT', |
| |
| ' will', 'will', ' willingly', |
| |
| 'In', ' In', |
| |
| ' new', ' newly', 'new', |
| |
| ' This', 'This', 'this', 'THIS', |
| |
| 've', "'ve", |
| |
| ' there', ' There', 'There', |
| |
| 'A', |
| |
| ' just', ' Just', 'Just', ' JUST', |
| |
| ' than', 'than', |
| |
| ' say', ' says', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'the-family', 1 |
|
|
| def prog_L06F02473(input_tokens): |
| """21 token families, 27 surface forms (train F1 of the uncapped optimum: 0.640). |
| Fires at 4,601 of 36,786,566 positions.""" |
| |
| """prev SAE description: the word "Ban" and its variations in various contexts""" |
| |
| fire = { |
| |
| ' ban', ' Ban', 'ban', 'Ban', ' Bans', |
| |
| ' banana', ' bananas', ' Banana', |
| |
| ' Barbara', |
| |
| ' Bri', |
| |
| ' Bild', |
| |
| ' Banner', |
| |
| ' Bac', |
| |
| ' Grab', |
| |
| ' Nurs', |
| |
| 'ande', |
| |
| 'Bang', |
| |
| 'anan', |
| |
| 'aban', |
| |
| ' Bok', |
| |
| ' Haunted', |
| |
| 'bara', |
| |
| ' Buk', |
| |
| ' Paran', |
| |
| 'Bomb', |
| |
| ' Fac', |
| |
| 'atu', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'ban-family', 1 |
|
|
| def prog_L06F02474(input_tokens): |
| """41 token families, 46 surface forms (train F1 of the uncapped optimum: 0.639). |
| Fires at 5,975 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the word "hol" indicating a focus on holidays or holiday-related topics""" |
| |
| fire = { |
| |
| 'hol', ' hol', 'Hol', 'holes', ' HOL', |
| |
| ' Housing', |
| |
| 'col', ' COL', |
| |
| ' battleground', |
| |
| 'hole', |
| |
| ' IL', |
| |
| ' Boyd', |
| |
| 'cut', |
| |
| ' Hut', |
| |
| 'hetic', |
| |
| 'hal', |
| |
| 'bur', |
| |
| 'othe', |
| |
| ' holog', |
| |
| 'lif', |
| |
| 'cerpt', |
| |
| 'hum', |
| |
| ' Brid', |
| |
| 'wal', |
| |
| ' jew', |
| |
| ' dispens', |
| |
| 'oven', |
| |
| 'oland', |
| |
| ' inhal', |
| |
| ' Lub', |
| |
| ' holistic', |
| |
| 'enthal', |
| |
| 'ohl', |
| |
| ' recol', |
| |
| 'ibia', |
| |
| ' holster', |
| |
| ' Hib', |
| |
| ' ..', |
| |
| ' doct', |
| |
| 'hod', |
| |
| ' discrim', |
| |
| 'aldehyde', |
| |
| ' elim', |
| |
| ' 808', |
| |
| 'inyl', |
| |
| ' weld', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'hol-family', 1 |
|
|
| def prog_L06F02475(input_tokens): |
| """5 token families, 5 surface forms (train F1 of the uncapped optimum: 0.259). |
| Fires at 52,742 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the letter 's'""" |
| |
| fire = { |
| |
| 's', |
| |
| 't', |
| |
| 're', |
| |
| 've', |
| |
| 'll', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 's-family', 1 |
|
|
| def prog_L06F02476(input_tokens): |
| """32 token families, 54 surface forms (train F1 of the uncapped optimum: 0.348). |
| Fires at 50,270 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms associated with damage, marks, or physical deterioration""" |
| |
| fire = { |
| |
| 'ed', |
| |
| 'y', |
| |
| 'red', ' red', |
| |
| ' burned', ' burns', ' burning', ' burn', 'burn', |
| |
| 'ashed', 'ashing', ' ash', 'ashes', ' ashes', |
| |
| 'ished', 'ishes', 'ishable', |
| |
| ' damaged', |
| |
| 'inging', 'ingly', 'inged', |
| |
| ' scar', ' scars', |
| |
| 'aced', |
| |
| 'ened', |
| |
| ' painted', 'painted', |
| |
| 'attered', 'attering', |
| |
| 'ted', |
| |
| 'aked', |
| |
| 'gy', |
| |
| 'amed', |
| |
| 'iled', |
| |
| 'iling', |
| |
| 'aming', |
| |
| 'rench', 'renched', |
| |
| 'ALT', 'alted', |
| |
| ' dirty', |
| |
| ' burnt', |
| |
| 'usted', |
| |
| ' sore', |
| |
| 'aint', 'ainted', |
| |
| ' battered', |
| |
| ' ble', |
| |
| ' freshly', 'Fresh', 'fresh', |
| |
| ' shattered', ' Shattered', |
| |
| 'ching', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'ed-family', 1 |
|
|
| def prog_L06F02477(input_tokens): |
| """31 token families, 57 surface forms (train F1 of the uncapped optimum: 0.296). |
| Fires at 85,035 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms related to medical conditions and biochemical processes""" |
| |
| fire = { |
| |
| ' b', |
| |
| ' p', |
| |
| ' l', |
| |
| ' m', |
| |
| ' micro', ' Micro', 'micro', 'Micro', |
| |
| ' non', ' nons', |
| |
| ' t', |
| |
| ' c', |
| |
| ' g', |
| |
| ' neuro', ' Neuro', |
| |
| ' o', |
| |
| ' ch', |
| |
| ' st', |
| |
| ' hyper', ' hyp', ' Hyp', ' hypers', 'hyper', |
| |
| ' tr', ' TR', |
| |
| ' sp', |
| |
| ' fl', |
| |
| ' sub', |
| |
| ' gl', ' Gl', |
| |
| ' breast', ' Breast', |
| |
| ' cl', |
| |
| ' par', 'Par', ' pars', |
| |
| ' inter', ' int', |
| |
| ' bio', ' bios', |
| |
| ' le', |
| |
| ' poly', ' Poly', 'poly', 'Poly', |
| |
| ' chronic', ' Chronic', |
| |
| ' trans', |
| |
| ' ar', |
| |
| ' mult', ' Mult', ' multic', 'Mult', 'mult', |
| |
| ' ac', ' Ac', 'Ac', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'b-family', 1 |
|
|
| def prog_L06F02478(input_tokens): |
| """48 token families, 53 surface forms (train F1 of the uncapped optimum: 0.593). |
| Fires at 10,483 of 36,786,566 positions.""" |
| |
| """prev SAE description: words related to the anatomical or physiological aspects of certain organisms or health topics""" |
| |
| fire = { |
| |
| 'rot', 'rots', 'Rot', |
| |
| ' Tayyip', |
| |
| 'pin', |
| |
| 'pert', |
| |
| 'bot', |
| |
| 'ERT', |
| |
| 'vert', 'VERT', |
| |
| 'chet', |
| |
| 'cop', |
| |
| ' tet', ' Tet', |
| |
| 'ullah', |
| |
| 'ilt', |
| |
| ' Recep', |
| |
| 'opt', 'opter', |
| |
| ' Rat', |
| |
| 'cript', |
| |
| ' staunch', |
| |
| ' RC', |
| |
| 'pel', |
| |
| 'arat', |
| |
| 'mot', |
| |
| 'rov', |
| |
| 'ledge', |
| |
| 'POL', |
| |
| 'rict', |
| |
| ' Prec', |
| |
| 'arson', |
| |
| 'lat', |
| |
| 'pit', |
| |
| 'OTUS', |
| |
| 'plot', |
| |
| 'Scient', |
| |
| 'rote', |
| |
| 'bent', |
| |
| 'roach', |
| |
| ' pancreat', |
| |
| 'rill', |
| |
| 'bey', |
| |
| ' neurot', |
| |
| ' antit', |
| |
| ' nont', |
| |
| 'rero', |
| |
| ' retard', |
| |
| 'achy', |
| |
| 'rosis', |
| |
| ' Panc', |
| |
| 'REC', |
| |
| ' relat', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'rot-family', 1 |
|
|
| def prog_L06F02479(input_tokens): |
| """38 token families, 54 surface forms (train F1 of the uncapped optimum: 0.677). |
| Fires at 6,862 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of graphs or graphical representations""" |
| |
| fire = { |
| |
| ' graph', ' graphical', ' graphs', ' Graph', 'Graph', 'graph', |
| |
| 'Abstract', |
| |
| 'Plot', 'plot', ' Plot', |
| |
| ' Matrix', ' matrix', 'Matrix', |
| |
| ' graphene', |
| |
| ' diagram', ' diagrams', |
| |
| ' charts', |
| |
| ' plotting', ' plotted', |
| |
| ' metaphor', ' metaphors', |
| |
| ' MRI', |
| |
| 'ogram', 'ograms', |
| |
| ' visualization', |
| |
| ' dom', |
| |
| ' Garn', ' garn', |
| |
| ' dynam', ' Dynam', |
| |
| ' quizz', |
| |
| ' axis', |
| |
| 'onz', |
| |
| 'raph', |
| |
| 'omorphic', 'omorph', |
| |
| ' dots', |
| |
| 'artz', |
| |
| ' lyr', |
| |
| ' crystall', |
| |
| ' mart', |
| |
| 'Ax', |
| |
| ' visualize', |
| |
| ' syrup', |
| |
| ' Rosenberg', |
| |
| ' hist', |
| |
| ' latt', |
| |
| 'ynam', |
| |
| ' veins', |
| |
| 'APH', |
| |
| ' Quartz', |
| |
| 'igraph', |
| |
| ' sarcast', |
| |
| 'stats', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'graph-family', 1 |
|
|
| def prog_L06F02480(input_tokens): |
| """10 token families, 12 surface forms (train F1 of the uncapped optimum: 0.123). |
| Fires at 828,321 of 36,786,566 positions. |
| Fires at 2.25% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: quantitative data and references to measurements or statistics""" |
| |
| fire = { |
| |
| ' the', |
| |
| ' of', |
| |
| ' in', |
| |
| ' for', |
| |
| ' with', |
| |
| ' from', |
| |
| ' new', 'new', ' newer', |
| |
| ' $', |
| |
| ' into', |
| |
| ' against', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'the-family', 1 |
|
|
| def prog_L06F02481(input_tokens): |
| """1 token families, 2 surface forms (train F1 of the uncapped optimum: 0.041). |
| Fires at 35,987 of 36,786,566 positions.""" |
| |
| """prev SAE description: repeated occurrences and emphasis on the word "the." """ |
| |
| fire = { |
| |
| ' the', 'the', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'the-family', 1 |
|
|
| def prog_L06F02482(input_tokens): |
| """9 token families, 11 surface forms (train F1 of the uncapped optimum: 0.075). |
| Fires at 190,605 of 36,786,566 positions. |
| Fires at 0.52% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: patterns of punctuation and formatting in the text, particularly focusing on the use of commas and other symbols""" |
| |
| fire = { |
| |
| ',', ' ,', |
| |
| ' (', |
| |
| '!', |
| |
| '),', ' ),', |
| |
| "'", |
| |
| ' –', |
| |
| ',"', |
| |
| ' --', |
| |
| '—', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return ',-family', 1 |
|
|
| def prog_L06F02483(input_tokens): |
| """10 token families, 14 surface forms (train F1 of the uncapped optimum: 0.075). |
| Fires at 84,604 of 36,786,566 positions.""" |
| |
| """prev SAE description: structural elements or punctuation commonly used in different contexts""" |
| |
| fire = { |
| |
| ' �', |
| |
| ' (', '(', |
| |
| ' [', |
| |
| ' very', |
| |
| ' Note', ' note', ' NOTE', |
| |
| ' self', 'self', |
| |
| ' kind', |
| |
| ' entire', |
| |
| ' #', |
| |
| ' email', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '�-family', 1 |
|
|
| def prog_L06F02484(input_tokens): |
| """12 token families, 19 surface forms (train F1 of the uncapped optimum: 0.811). |
| Fires at 9,821 of 36,786,566 positions.""" |
| |
| """prev SAE description: the word "General" in various contexts""" |
| |
| fire = { |
| |
| ' General', ' general', 'General', 'general', ' GENERAL', |
| |
| ' Gen', ' Gener', 'GEN', |
| |
| ' Colonel', ' colonel', |
| |
| ' Admiral', |
| |
| ' Gene', |
| |
| ' generals', |
| |
| ' Marshal', |
| |
| ' Gregg', |
| |
| ' Hampton', |
| |
| 'onel', |
| |
| ' constituent', |
| |
| ' Supplemental', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'gener-family', 1 |
|
|
| def prog_L06F02485(input_tokens): |
| """21 token families, 32 surface forms (train F1 of the uncapped optimum: 0.254). |
| Fires at 17,836 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to external reading or additional resources""" |
| |
| fire = { |
| |
| ' read', ' reading', ' Read', ' Reading', 'Read', 'READ', ' READ', |
| |
| 'IVE', |
| |
| 'vertising', |
| |
| 'Activity', |
| |
| 'RELATED', 'Related', |
| |
| ' Articles', |
| |
| 'VERTIS', |
| |
| 'OST', |
| |
| 'lights', |
| |
| ' ALSO', |
| |
| ' Know', ' KNOW', |
| |
| ' Picks', |
| |
| ' EST', |
| |
| ' Videos', ' VIDEOS', |
| |
| ' Stories', ' STOR', |
| |
| ' RELEASE', |
| |
| 'SEE', ' SEE', |
| |
| ' LAT', |
| |
| ' Reasons', |
| |
| ' Newsletter', |
| |
| ' Content', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'read-family', 1 |
|
|
| def prog_L06F02486(input_tokens): |
| """39 token families, 40 surface forms (train F1 of the uncapped optimum: 0.814). |
| Fires at 7,153 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to the name "Justin." """ |
| |
| fire = { |
| |
| ' Justin', 'Justin', |
| |
| ' Evan', |
| |
| ' Sebastian', |
| |
| ' Zuckerberg', |
| |
| ' Devin', |
| |
| ' Cody', |
| |
| ' selfie', |
| |
| ' Bieber', |
| |
| ' Shia', |
| |
| ' lemon', |
| |
| ' Geek', |
| |
| ' Malik', |
| |
| ' Whitney', |
| |
| ' Shinzo', |
| |
| 'Joshua', |
| |
| ' Winston', |
| |
| ' Zhang', |
| |
| ' Sophie', |
| |
| ' selfies', |
| |
| ' Manziel', |
| |
| ' GoPro', |
| |
| 'aden', |
| |
| ' Stefan', |
| |
| ' Kimberly', |
| |
| ' Millennials', |
| |
| ' Fac', |
| |
| ' Juno', |
| |
| ' Sophia', |
| |
| 'udeau', |
| |
| ' Johann', |
| |
| 'evin', |
| |
| 'phen', |
| |
| ' twe', |
| |
| ' influ', |
| |
| ' Jian', |
| |
| ' Sina', |
| |
| 'fruit', |
| |
| ' Millenn', |
| |
| ' Clin', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'justin-family', 1 |
|
|
| def prog_L06F02487(input_tokens): |
| """19 token families, 62 surface forms (train F1 of the uncapped optimum: 0.186). |
| Fires at 330,061 of 36,786,566 positions. |
| Fires at 0.90% of positions, and |
| token rules do not serve features this dense: median test F1 0.086 |
| across the 37,173 features above 0.39% of positions, against 0.763 for |
| the 76,283 below 0.016%. Read this fire set as the best available |
| approximation, not as what the feature does.""" |
| |
| """prev SAE description: phrases related to viral internet trends or sensations""" |
| |
| fire = { |
| |
| ' on', ' ON', |
| |
| ' online', ' Online', 'online', |
| |
| ' quickly', ' quick', |
| |
| ' market', ' markets', ' Market', 'Market', ' Markets', |
| |
| ' world', ' worlds', ' WORLD', |
| |
| ' where', 'where', 'WHERE', ' WHERE', |
| |
| ' Internet', ' internet', |
| |
| ' growing', 'growing', ' grow', ' grows', |
| |
| ' Bitcoin', ' bitcoin', 'Bitcoin', ' Bitcoins', 'bitcoin', |
| |
| ' site', |
| |
| ' leading', 'leader', ' Leading', |
| |
| ' social', |
| |
| ' shares', ' Shares', ' shared', ' sharing', 'Shares', 'sharing', 'shared', |
| |
| ' hot', 'hot', ' hotly', ' Hot', ' HOT', |
| |
| ' stock', ' Stock', ' stocks', 'stock', |
| |
| ' place', |
| |
| ' #', |
| |
| ' search', ' searching', ' Search', ' searched', ' searches', 'search', |
| |
| ' hit', ' hits', ' Hits', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'on-family', 1 |
|
|
| def prog_L06F02488(input_tokens): |
| """50 token families, 52 surface forms (train F1 of the uncapped optimum: 0.787). |
| Fires at 8,151 of 36,786,566 positions.""" |
| |
| """prev SAE description: mentions of the term "Auburn" and related references to Auburn University""" |
| |
| fire = { |
| |
| ' Tigers', |
| |
| ' Troy', |
| |
| ' Albany', |
| |
| ' Auburn', |
| |
| ' Zimbabwe', |
| |
| ' Azure', |
| |
| ' LSU', |
| |
| ' Fut', |
| |
| ' Kabul', |
| |
| ' Aston', |
| |
| ' Clemson', |
| |
| ' Syracuse', |
| |
| ' Arbor', |
| |
| 'aux', |
| |
| ' Bulldogs', |
| |
| ' Gators', |
| |
| ' Vine', |
| |
| ' bast', |
| |
| ' Vanderbilt', |
| |
| ' Bentley', |
| |
| ' Carlton', |
| |
| ' Azerbaijan', |
| |
| 'endon', |
| |
| ' Fresno', |
| |
| ' swamp', |
| |
| ' Vale', |
| |
| ' Yak', |
| |
| ' Aus', |
| |
| ' Homer', |
| |
| ' AU', |
| |
| ' Mug', |
| |
| ' Beacon', |
| |
| ' Aero', |
| |
| ' Bellev', |
| |
| ' Augusta', |
| |
| ' Aub', |
| |
| 'UA', ' UA', |
| |
| ' Vulcan', |
| |
| 'mson', |
| |
| ' Wak', |
| |
| ' Apex', |
| |
| ' grounding', |
| |
| ' agility', ' Agility', |
| |
| 'igers', |
| |
| ' Oblivion', |
| |
| ' ACE', |
| |
| 'imbabwe', |
| |
| ' Ascend', |
| |
| ' AX', |
| |
| ' Ake', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'tig-family', 1 |
|
|
| def prog_L06F02489(input_tokens): |
| """1 token families, 2 surface forms (train F1 of the uncapped optimum: 0.076). |
| Fires at 59,146 of 36,786,566 positions.""" |
| |
| """prev SAE description: phrases related to the results of studies or research""" |
| |
| fire = { |
| |
| ' of', ' OF', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'of-family', 1 |
|
|
| def prog_L06F02490(input_tokens): |
| """45 token families, 59 surface forms (train F1 of the uncapped optimum: 0.761). |
| Fires at 9,001 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the word "Wat" in various contexts""" |
| |
| fire = { |
| |
| ' Wat', ' wat', 'wat', |
| |
| 'wen', ' Wen', |
| |
| ' Hos', |
| |
| ' Ot', |
| |
| ' Wake', |
| |
| ' Wy', |
| |
| ' Fut', |
| |
| 'won', ' Won', |
| |
| ' Kob', |
| |
| ' Ow', ' OW', |
| |
| ' Hornets', |
| |
| ' Wolves', |
| |
| ' Wan', 'Wan', |
| |
| ' Hut', ' hut', |
| |
| ' Saw', |
| |
| ' Wer', |
| |
| ' mim', |
| |
| ' onlook', |
| |
| 'cop', |
| |
| ' Watkins', |
| |
| ' Iw', |
| |
| ' Wit', 'wit', |
| |
| 'prev', ' Prev', |
| |
| ' swat', |
| |
| ' Tut', |
| |
| ' wond', ' Wond', |
| |
| ' worsh', |
| |
| ' Hin', |
| |
| ' Sear', ' sear', |
| |
| ' Myr', |
| |
| ' Toby', |
| |
| ' WT', 'wt', |
| |
| 'kept', |
| |
| ' HS', |
| |
| ' Wak', |
| |
| ' Sut', |
| |
| ' Viz', |
| |
| 'gat', |
| |
| ' Farn', |
| |
| ' widgets', 'widget', 'Widget', |
| |
| ' Gon', |
| |
| ' Wad', |
| |
| 'idon', |
| |
| ' Owl', |
| |
| ' XT', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'wat-family', 1 |
|
|
| def prog_L06F02491(input_tokens): |
| """34 token families, 61 surface forms (train F1 of the uncapped optimum: 0.309). |
| Fires at 141,881 of 36,786,566 positions.""" |
| |
| """prev SAE description: verbs related to construction, development, and legislative actions""" |
| |
| fire = { |
| |
| ' make', |
| |
| ' allow', ' allowing', 'allow', ' Allow', |
| |
| ' bring', ' Bring', 'bring', |
| |
| ' build', 'build', |
| |
| ' protect', ' protecting', ' Protect', 'Protect', |
| |
| ' keep', |
| |
| ' create', |
| |
| ' replace', ' Replace', |
| |
| ' provide', ' Provide', |
| |
| ' ban', ' bans', |
| |
| ' raise', ' Raise', |
| |
| ' reduce', ' Reduce', |
| |
| ' remove', ' Remove', 'remove', |
| |
| ' improve', ' Improve', 'improve', |
| |
| ' add', |
| |
| ' hold', |
| |
| ' save', ' Save', 'save', |
| |
| ' prevent', ' Prevent', |
| |
| ' ensure', ' Ensure', |
| |
| ' expand', |
| |
| ' develop', |
| |
| ' receive', |
| |
| ' sell', |
| |
| ' send', |
| |
| ' join', 'join', |
| |
| ' pass', |
| |
| ' establish', ' establishing', 'establish', |
| |
| ' repeal', |
| |
| 'ize', |
| |
| ' extend', ' extending', ' Extend', |
| |
| ' promote', |
| |
| ' defend', |
| |
| ' investigate', |
| |
| ' celebrate', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'make-family', 1 |
|
|
| def prog_L06F02492(input_tokens): |
| """10 token families, 10 surface forms (train F1 of the uncapped optimum: 0.739). |
| Fires at 4,483 of 36,786,566 positions.""" |
| |
| """prev SAE description: occurrences of the letters "ch" in various contexts""" |
| |
| fire = { |
| |
| ' ch', |
| |
| ' sch', |
| |
| ' lat', |
| |
| ' chim', |
| |
| ' champ', |
| |
| ' gon', |
| |
| ' chimpanzees', |
| |
| ' chimpan', |
| |
| ' phil', |
| |
| 'ileged', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'ch-family', 1 |
|
|
| def prog_L06F02493(input_tokens): |
| """34 token families, 58 surface forms (train F1 of the uncapped optimum: 0.225). |
| Fires at 27,005 of 36,786,566 positions.""" |
| |
| """prev SAE description: verbs related to negative actions or criticism""" |
| |
| fire = { |
| |
| 'ing', 'ING', 'inging', |
| |
| 'ashed', 'ashing', 'ashes', |
| |
| 'ENT', 'ented', 'ental', 'enting', |
| |
| 'acking', 'acked', |
| |
| 'IA', |
| |
| 'apper', 'apping', 'apped', |
| |
| 'ises', 'ISE', |
| |
| 'lines', |
| |
| 'ests', |
| |
| 'angers', 'anging', 'anges', |
| |
| 'aging', |
| |
| 'ailed', 'ailing', |
| |
| 'ift', 'ifts', 'ifting', 'ifter', 'IFT', |
| |
| 'ising', |
| |
| 'aged', |
| |
| 'ble', |
| |
| 'ages', |
| |
| 'anking', 'anks', 'ANK', 'anked', |
| |
| 'ough', 'OUGH', |
| |
| 'USH', 'ushing', |
| |
| 'OME', |
| |
| 'iled', |
| |
| 'ving', |
| |
| 'auc', |
| |
| ' rave', |
| |
| 'ammed', |
| |
| 'esc', |
| |
| 'acing', |
| |
| 'uling', |
| |
| 'ips', |
| |
| 'ousing', 'oused', |
| |
| 'iding', |
| |
| 'iddle', |
| |
| 'ums', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'ing-family', 1 |
|
|
| def prog_L06F02494(input_tokens): |
| """46 token families, 55 surface forms (train F1 of the uncapped optimum: 0.685). |
| Fires at 9,652 of 36,786,566 positions.""" |
| |
| """prev SAE description: patterns related to the term "urn." """ |
| |
| fire = { |
| |
| 'urn', 'URN', 'urned', |
| |
| ' Turnbull', |
| |
| 'burn', 'Burn', |
| |
| 'UST', |
| |
| ' Horn', |
| |
| ' cough', ' coughing', |
| |
| ' Whedon', |
| |
| 'ither', |
| |
| ' Saturn', |
| |
| 'utton', |
| |
| ' Coff', ' coff', |
| |
| ' AOL', |
| |
| 'tain', |
| |
| ' Homer', |
| |
| ' Reign', |
| |
| ' Hoover', |
| |
| ' heterosexual', |
| |
| ' farewell', |
| |
| 'pload', |
| |
| ' Alban', |
| |
| 'rend', |
| |
| 'aign', |
| |
| '370', |
| |
| ' conven', |
| |
| ' tarn', |
| |
| ' Requ', 'Requ', ' requ', |
| |
| 'olith', |
| |
| ' Graves', |
| |
| ' Furn', ' furn', |
| |
| 'ORPG', |
| |
| ' Morrow', |
| |
| ' gown', |
| |
| ' cavern', |
| |
| ' falsehood', |
| |
| ' Garn', ' garn', |
| |
| ' monopol', |
| |
| ' Firefly', |
| |
| 'orrow', |
| |
| ' Nirvana', |
| |
| ' Roof', |
| |
| ' Uniform', |
| |
| ' shaft', |
| |
| 'iform', |
| |
| 'houn', |
| |
| ' Rao', |
| |
| 'egg', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'urn-family', 1 |
|
|
| def prog_L06F02495(input_tokens): |
| """34 token families, 51 surface forms (train F1 of the uncapped optimum: 0.178). |
| Fires at 55,329 of 36,786,566 positions.""" |
| |
| """prev SAE description: terms related to foreclosure and financial distress""" |
| |
| fire = { |
| |
| '-', |
| |
| 'v', 'V', |
| |
| 'l', |
| |
| 'b', |
| |
| ' Images', 'Images', |
| |
| 'st', |
| |
| 'n', |
| |
| 'c', |
| |
| 'am', |
| |
| 'f', |
| |
| 'g', |
| |
| 'k', |
| |
| 'ANC', |
| |
| 'av', 'AV', 'Av', |
| |
| 'VEL', 'vel', |
| |
| 'id', 'ID', |
| |
| 'amb', |
| |
| 'TH', |
| |
| 'w', |
| |
| 'r', |
| |
| ' Dame', |
| |
| 'ING', |
| |
| 'az', |
| |
| 'cl', |
| |
| ' Korea', |
| |
| 'qu', 'QU', |
| |
| 'cast', 'casters', 'caster', 'casting', 'casts', |
| |
| 'ana', 'anas', |
| |
| ' Reader', ' Readers', 'reads', |
| |
| 'VER', 'ver', 'VERS', |
| |
| 'od', |
| |
| 'ater', 'ATER', |
| |
| 'UT', |
| |
| 'iv', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return '--family', 1 |
|
|
| def prog_L06F02496(input_tokens): |
| """29 token families, 42 surface forms (train F1 of the uncapped optimum: 0.572). |
| Fires at 11,167 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to various types of forms""" |
| |
| fire = { |
| |
| ' form', ' forms', 'form', ' Form', 'Form', 'FORM', ' Forms', ' FORM', |
| |
| ' Fill', 'Fill', |
| |
| ' paperwork', |
| |
| 'edin', |
| |
| ' broker', ' brokers', |
| |
| 'Disclaimer', |
| |
| ' Craigslist', |
| |
| ' calculator', ' Calculator', |
| |
| ' fraudulent', |
| |
| ' questionnaire', |
| |
| 'METHOD', 'Methods', ' Methods', |
| |
| ' Excel', |
| |
| ' envelope', |
| |
| 'omin', |
| |
| ' spreadsheet', |
| |
| ' FEC', |
| |
| 'itte', |
| |
| ' envelop', |
| |
| ' Coinbase', |
| |
| ' broch', |
| |
| 'sheets', |
| |
| ' Tinder', |
| |
| 'osc', |
| |
| ' summons', |
| |
| ' Stras', |
| |
| 'ifax', |
| |
| ' masse', |
| |
| 'upid', |
| |
| ' fax', 'Fax', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'form-family', 1 |
|
|
| def prog_L06F02497(input_tokens): |
| """19 token families, 23 surface forms (train F1 of the uncapped optimum: 0.397). |
| Fires at 66,785 of 36,786,566 positions.""" |
| |
| """prev SAE description: proper names and notable figures in sports and entertainment""" |
| |
| fire = { |
| |
| 'man', |
| |
| 'berg', 'berger', |
| |
| 'mann', ' Manning', |
| |
| 'ley', |
| |
| 'ner', |
| |
| 'ino', |
| |
| 'ke', |
| |
| 'Reilly', |
| |
| ' Brady', |
| |
| 'kin', 'kins', |
| |
| ' Ford', |
| |
| 'ovic', |
| |
| ' Blasio', |
| |
| ' Gaal', |
| |
| 'erson', |
| |
| ' Sox', |
| |
| 'kowski', |
| |
| ' Wings', 'Wing', |
| |
| 'key', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'man-family', 1 |
|
|
| def prog_L06F02498(input_tokens): |
| """3 token families, 5 surface forms (train F1 of the uncapped optimum: 0.385). |
| Fires at 12,381 of 36,786,566 positions.""" |
| |
| """prev SAE description: expressions related to having a good or great time""" |
| |
| fire = { |
| |
| ' time', 'time', ' TIME', |
| |
| ' Sonny', |
| |
| 'tim', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'time-family', 1 |
|
|
| def prog_L06F02499(input_tokens): |
| """8 token families, 11 surface forms (train F1 of the uncapped optimum: 0.601). |
| Fires at 8,951 of 36,786,566 positions.""" |
| |
| """prev SAE description: references to various human needs""" |
| |
| fire = { |
| |
| ' needs', ' Needs', 'needs', ' NEED', |
| |
| ' requirements', |
| |
| ' Achilles', |
| |
| ' necessities', |
| |
| ' needy', |
| |
| ' desires', |
| |
| ' inadequ', |
| |
| ' Kear', |
| } |
| token = input_tokens[-1] |
| if token not in fire: |
| return "no match", 0 |
| return 'need-family', 1 |
|
|
