[ { "name": "shared_bpe_8000", "tokenizer_type": "shared", "algorithm": "bpe", "vocab_size": 8000, "fertility_overall": 1.5744684296670828, "fertility_ar": 1.4755656463463651, "fertility_az": 1.7595859759130557, "cpt_overall": 3.5313313486070697, "cpt_ar": 3.812953050264278, "cpt_az": 3.0042165735724513, "fertility_disparity": 0.16141315823985886, "cpt_disparity": 0.21210239571026782, "oov_rate": 0.0, "vocab_gini": 0.7419718947527658, "shannon_entropy": 10.725417379440934, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_bpe_8000", "tokenizer_type": "concatenated", "algorithm": "bpe", "vocab_size": 8000, "fertility_overall": 1.625089409755206, "fertility_ar": 1.677255929872786, "fertility_az": 1.5274486975717831, "cpt_overall": 3.4121822358457132, "cpt_ar": 3.3745329363977343, "cpt_az": 3.4826508889784487, "fertility_disparity": 0.0893168595399542, "cpt_disparity": 0.031044728865266227, "oov_rate": 0.0, "vocab_gini": 0.7155357357913302, "shannon_entropy": 10.758976129080413, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 }, { "name": "shared_unigram_8000", "tokenizer_type": "shared", "algorithm": "unigram", "vocab_size": 8000, "fertility_overall": 1.5860234519252308, "fertility_ar": 1.5098217608254827, "fertility_az": 1.7286510868870875, "cpt_overall": 3.513922137310213, "cpt_ar": 3.7495598617218993, "cpt_az": 3.072876133532698, "fertility_disparity": 0.12658964421540223, "cpt_disparity": 0.18047017600579116, "oov_rate": 0.0, "vocab_gini": 0.8011276471980863, "shannon_entropy": 10.278366552533573, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_unigram_8000", "tokenizer_type": "concatenated", "algorithm": "unigram", "vocab_size": 8000, "fertility_overall": 1.630557495524919, "fertility_ar": 1.6828221737842166, "fertility_az": 1.5327330595500248, "cpt_overall": 3.4256065103693167, "cpt_ar": 3.3890978047599254, "cpt_az": 3.4939403000980764, "fertility_disparity": 0.08918893307465849, "cpt_disparity": 0.030006950987459077, "oov_rate": 0.0, "vocab_gini": 0.7843741193591066, "shannon_entropy": 10.258501048163726, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 }, { "name": "shared_wordpiece_8000", "tokenizer_type": "shared", "algorithm": "wordpiece", "vocab_size": 8000, "fertility_overall": 1.5720141231575526, "fertility_ar": 1.4715587661386824, "fertility_az": 1.7600376405371791, "cpt_overall": 3.5349251084819273, "cpt_ar": 3.819823570325556, "cpt_az": 3.001677181426908, "fertility_disparity": 0.16390494598197936, "cpt_disparity": 0.21418433962616734, "oov_rate": 0.0, "vocab_gini": 0.7254354187229622, "shannon_entropy": 10.734803471006908, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "concat_wordpiece_8000", "tokenizer_type": "concatenated", "algorithm": "wordpiece", "vocab_size": 8000, "fertility_overall": 1.6305196797375283, "fertility_ar": 1.679880351462672, "fertility_az": 1.5381307082241507, "cpt_overall": 3.397665609111008, "cpt_ar": 3.366673383613362, "cpt_az": 3.4556741357085015, "fertility_disparity": 0.08438079718897826, "cpt_disparity": 0.02575496085567456, "oov_rate": 0.0, "vocab_gini": 0.6940886180639433, "shannon_entropy": 10.748321790799139, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "shared_bbpe_8000", "tokenizer_type": "shared", "algorithm": "bbpe", "vocab_size": 8000, "fertility_overall": 1.7155041180862605, "fertility_ar": 1.5432842580927282, "fertility_az": 2.037850131276259, "cpt_overall": 3.2675384671739276, "cpt_ar": 3.6389791874263038, "cpt_az": 2.5723083391604016, "fertility_disparity": 0.24269001218151182, "cpt_disparity": 0.2931236463103581, "oov_rate": 0.0, "vocab_gini": 0.7852085662214892, "shannon_entropy": 10.17036760209476, "exact_match_ar": 0.9994106980961015, "exact_match_az": 0.9994909214322077 }, { "name": "concat_bbpe_8000", "tokenizer_type": "concatenated", "algorithm": "bbpe", "vocab_size": 8000, "fertility_overall": 1.7695328842122233, "fertility_ar": 1.733122664722859, "fertility_az": 1.8376823362685084, "cpt_overall": 3.130408968636395, "cpt_ar": 3.268439481034545, "cpt_az": 2.872055574482044, "fertility_disparity": 0.056897576627940066, "cpt_disparity": 0.12127619582756817, "oov_rate": 0.0, "vocab_gini": 0.7652784949485903, "shannon_entropy": 10.232825855115562, "exact_match_ar": 0.999909338168631, "exact_match_az": 0.9997454607161038 }, { "name": "shared_bpe_16000", "tokenizer_type": "shared", "algorithm": "bpe", "vocab_size": 16000, "fertility_overall": 1.406466326610023, "fertility_ar": 1.3322457895485136, "fertility_az": 1.545385811386613, "cpt_overall": 3.9381818445330614, "cpt_ar": 4.211736441809671, "cpt_az": 3.426166367193676, "fertility_disparity": 0.13792026577936378, "cpt_disparity": 0.18651928615895444, "oov_rate": 0.0, "vocab_gini": 0.7705612275154623, "shannon_entropy": 11.22051844286477, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_bpe_16000", "tokenizer_type": "concatenated", "algorithm": "bpe", "vocab_size": 16000, "fertility_overall": 1.4427306790554728, "fertility_ar": 1.4863282848922221, "fertility_az": 1.3611285082800872, "cpt_overall": 3.843167233342444, "cpt_ar": 3.806560946253488, "cpt_az": 3.9116836675170874, "fertility_disparity": 0.08423426902705648, "cpt_disparity": 0.02687403435419539, "oov_rate": 0.0, "vocab_gini": 0.7502784093339115, "shannon_entropy": 11.289933167496804, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 }, { "name": "shared_unigram_16000", "tokenizer_type": "shared", "algorithm": "unigram", "vocab_size": 16000, "fertility_overall": 1.43915848471923, "fertility_ar": 1.383798509647741, "fertility_az": 1.542776425502961, "cpt_overall": 3.856083364492832, "cpt_ar": 4.074326867754011, "cpt_az": 3.4475943366681565, "fertility_disparity": 0.10304663282846806, "cpt_disparity": 0.1538248038089647, "oov_rate": 0.0, "vocab_gini": 0.8202410337846489, "shannon_entropy": 10.661726018051654, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_unigram_16000", "tokenizer_type": "concatenated", "algorithm": "unigram", "vocab_size": 16000, "fertility_overall": 1.4690109341605053, "fertility_ar": 1.5133852390561142, "fertility_az": 1.3859550062802124, "cpt_overall": 3.7990700186667543, "cpt_ar": 3.7636569223952008, "cpt_az": 3.865353143030446, "fertility_disparity": 0.08420211158883709, "cpt_disparity": 0.026309684231209756, "oov_rate": 0.0, "vocab_gini": 0.8131049651034313, "shannon_entropy": 10.680674921480614, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 }, { "name": "shared_wordpiece_16000", "tokenizer_type": "shared", "algorithm": "wordpiece", "vocab_size": 16000, "fertility_overall": 1.4017223613822598, "fertility_ar": 1.3278822099406427, "fertility_az": 1.5399298737530456, "cpt_overall": 3.949283003677344, "cpt_ar": 4.223103615213092, "cpt_az": 3.43676962420351, "fertility_disparity": 0.1376995585491242, "cpt_disparity": 0.18619812883040157, "oov_rate": 0.0, "vocab_gini": 0.7570477677642601, "shannon_entropy": 11.225647699841687, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "concat_wordpiece_16000", "tokenizer_type": "concatenated", "algorithm": "wordpiece", "vocab_size": 16000, "fertility_overall": 1.441782460775622, "fertility_ar": 1.4835885687430344, "fertility_az": 1.3635334584201908, "cpt_overall": 3.8430675831500807, "cpt_ar": 3.81082194270718, "cpt_az": 3.9034221417934196, "fertility_disparity": 0.08092210526032824, "cpt_disparity": 0.0237228246708911, "oov_rate": 0.0, "vocab_gini": 0.7327455145352297, "shannon_entropy": 11.280193007623335, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "shared_bbpe_16000", "tokenizer_type": "shared", "algorithm": "bbpe", "vocab_size": 16000, "fertility_overall": 1.563251758755881, "fertility_ar": 1.407745203379972, "fertility_az": 1.8543152757754426, "cpt_overall": 3.5796704536935944, "cpt_ar": 3.9802348152214964, "cpt_az": 2.8299292509695566, "fertility_disparity": 0.24082747860054313, "cpt_disparity": 0.2890044476403401, "oov_rate": 0.0, "vocab_gini": 0.8152607254871892, "shannon_entropy": 10.485685853996797, "exact_match_ar": 0.9994106980961015, "exact_match_az": 0.9994909214322077 }, { "name": "concat_bbpe_16000", "tokenizer_type": "concatenated", "algorithm": "bbpe", "vocab_size": 16000, "fertility_overall": 1.6044565681716603, "fertility_ar": 1.5508741409303313, "fertility_az": 1.7047474509939684, "cpt_overall": 3.4578261987966146, "cpt_ar": 3.653251881231358, "cpt_az": 3.0920455646111016, "fertility_disparity": 0.09026164548532975, "cpt_disparity": 0.15361829265139457, "oov_rate": 0.0, "vocab_gini": 0.8039903620272112, "shannon_entropy": 10.550324007405273, "exact_match_ar": 0.999909338168631, "exact_match_az": 0.9997454607161038 }, { "name": "shared_bpe_32000", "tokenizer_type": "shared", "algorithm": "bpe", "vocab_size": 32000, "fertility_overall": 1.2796960189731112, "fertility_ar": 1.2326026481342502, "fertility_az": 1.3678412557544855, "cpt_overall": 4.304399960106754, "cpt_ar": 4.537455476649384, "cpt_az": 3.8681871063030555, "fertility_disparity": 0.09887010429849855, "cpt_disparity": 0.1474986088107117, "oov_rate": 0.0, "vocab_gini": 0.7839455718558345, "shannon_entropy": 11.593993601001019, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_bpe_32000", "tokenizer_type": "concatenated", "algorithm": "bpe", "vocab_size": 32000, "fertility_overall": 1.2695706222898995, "fertility_ar": 1.3077466165859795, "fertility_az": 1.1981161479838143, "cpt_overall": 4.4229147884371125, "cpt_ar": 4.388664556275568, "cpt_az": 4.487021365858093, "fertility_disparity": 0.08383158267185423, "cpt_disparity": 0.021920290001497422, "oov_rate": 0.0, "vocab_gini": 0.7084770707714895, "shannon_entropy": 12.560099600380696, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "shared_unigram_32000", "tokenizer_type": "shared", "algorithm": "unigram", "vocab_size": 32000, "fertility_overall": 1.3340159750699838, "fertility_ar": 1.3027084505800648, "fertility_az": 1.3926146506382522, "cpt_overall": 4.136639454094978, "cpt_ar": 4.308287618942264, "cpt_az": 3.8153634896854167, "fertility_disparity": 0.06455928064305677, "cpt_disparity": 0.11441300415729108, "oov_rate": 0.0, "vocab_gini": 0.8250969225299254, "shannon_entropy": 10.965056609467814, "exact_match_ar": 0.9990480507706255, "exact_match_az": 0.9992363821483116 }, { "name": "concat_unigram_32000", "tokenizer_type": "concatenated", "algorithm": "unigram", "vocab_size": 32000, "fertility_overall": 1.3498732692189985, "fertility_ar": 1.392269196943319, "fertility_az": 1.2705202940282205, "cpt_overall": 4.117905456086874, "cpt_ar": 4.071549633983907, "cpt_az": 4.204670213900507, "fertility_disparity": 0.08744638119007023, "cpt_disparity": 0.03166017146279571, "oov_rate": 0.0, "vocab_gini": 0.8281405210808994, "shannon_entropy": 11.002763195930038, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 }, { "name": "shared_wordpiece_32000", "tokenizer_type": "shared", "algorithm": "wordpiece", "vocab_size": 32000, "fertility_overall": 1.2742505256521945, "fertility_ar": 1.2274699882356, "fertility_az": 1.3618102283002589, "cpt_overall": 4.319881028657028, "cpt_ar": 4.553239284995183, "cpt_az": 3.883101533084343, "fertility_disparity": 0.0986482824646834, "cpt_disparity": 0.1471782416793298, "oov_rate": 0.0, "vocab_gini": 0.7747440551637088, "shannon_entropy": 11.589595542112136, "exact_match_ar": 0.0, "exact_match_az": 0.0 }, { "name": "concat_wordpiece_32000", "tokenizer_type": "concatenated", "algorithm": "wordpiece", "vocab_size": 32000, "fertility_overall": 1.303008786024938, "fertility_ar": 1.3416327230816638, "fertility_az": 1.2307158917884393, "cpt_overall": 4.240397119793184, "cpt_ar": 4.201360994624003, "cpt_az": 4.313461511548837, "fertility_disparity": 0.08267302174804893, "cpt_disparity": 0.025988528383688383, "oov_rate": 0.0, "vocab_gini": 0.7687350341879174, "shannon_entropy": 11.65382526739417, "exact_match_ar": 0.001042611060743427, "exact_match_az": 0.00033938571186153065 }, { "name": "shared_bbpe_32000", "tokenizer_type": "shared", "algorithm": "bbpe", "vocab_size": 32000, "fertility_overall": 1.4537480484685892, "fertility_ar": 1.317581832517785, "fertility_az": 1.7086120162163183, "cpt_overall": 3.834237272964042, "cpt_ar": 4.240919303648946, "cpt_az": 3.0730455542376744, "fertility_disparity": 0.22885838328847793, "cpt_disparity": 0.2753822145133525, "oov_rate": 0.0, "vocab_gini": 0.8293407455122993, "shannon_entropy": 10.685544129991571, "exact_match_ar": 0.9994106980961015, "exact_match_az": 0.9994909214322077 }, { "name": "concat_bbpe_32000", "tokenizer_type": "concatenated", "algorithm": "bbpe", "vocab_size": 32000, "fertility_overall": 1.306778190060823, "fertility_ar": 1.3463628334026025, "fertility_az": 1.2326871301490916, "cpt_overall": 4.231787559188005, "cpt_ar": 4.190301018162676, "cpt_az": 4.3094384241989285, "fertility_disparity": 0.08443170030639024, "cpt_disparity": 0.02764569169088408, "oov_rate": 0.0, "vocab_gini": 0.7818011980546499, "shannon_entropy": 11.6730644846771, "exact_match_ar": 0.9989120580235721, "exact_match_az": 0.9964364500254539 } ]