{"meta":{"query_hash":"9a3c2b27b4d5","filters":{"venue":"Mining Software Repositories"},"cohort_total":17,"direct_labels_cover":0,"predictions_cover":17,"exported":17,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/9a3c2b27b4d5","api":"https://metacan.xera.ac/api/v1/cohort?venue=Mining+Software+Repositories"},"results":[{"id":"W1968137197","doi":"10.5555/2664446.2664466","title":"The build dependency perspective of Android's concrete architecture","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Advanced Software Engineering Methodologies","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Android (operating system); Architecture; Computer science; Android application; Applications architecture; Embedded system; Operating system; Software architecture; Software","score_opus":0.016041832489007275,"score_gpt":0.276711630110583,"score_spread":0.2606697976215757,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1968137197","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.10060001,0.00032345866,0.8729194,0.00076603494,0.000035837653,0.00021815186,0.000914344,0.0013002123,0.022922464],"genre_scores_gemma":[0.48001122,0.00068694656,0.5075984,0.00015624381,0.00003164797,0.00033036587,0.0024757106,0.00055716315,0.008152213],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.99802727,0.000442197,0.00020805789,0.00038304212,0.0007634969,0.00017593584],"domain_scores_gemma":[0.99566805,0.0017515708,0.0005064136,0.0011555695,0.0007917331,0.0001266739],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013707177,0.0006745109,0.00025734506,0.0028694826,0.0009544571,0.002423648,0.0007482212,0.0010827549,0.0032080396],"category_scores_gemma":[0.007215828,0.0010962359,0.0012534645,0.0019499451,0.0020826685,0.004930222,0.0015898902,0.0018657864,0.0007906544],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013684927,0.00014869105,0.041077163,0.00069037953,0.00016036021,0.0030358427,0.013799518,0.040742256,0.027357537,0.72470117,0.006081424,0.14206882],"study_design_scores_gemma":[0.000049192462,0.00027254614,0.052097425,0.00070885225,0.0004716942,0.0052174535,0.0046530627,0.25711003,0.043929383,0.35192737,0.28336513,0.00019784465],"about_ca_topic_score_codex":0.012438021,"about_ca_topic_score_gemma":0.01792985,"teacher_disagreement_score":0.012438021,"about_ca_system_score_codex":0.0013746682,"about_ca_system_score_gemma":0.002018601,"threshold_uncertainty_score":0.024731278},"labels":[],"label_agreement":null},{"id":"W1978832807","doi":"10.5555/2820518.2820554","title":"An empirical study of end-user programmers in the computer music community","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Metadata; Software; World Wide Web; Software development; Population; Source code; Computer programming; Musical; End user; Empirical research; Software engineering; Programming language","score_opus":0.0705591971718184,"score_gpt":0.3386892397372468,"score_spread":0.2681300425654284,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1978832807","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99903584,0.000056969984,0.00017284851,0.00012349687,0.000001863625,0.000026038397,0.000020053883,0.00000412603,0.00055863237],"genre_scores_gemma":[0.9982899,0.00019009669,0.00065901823,0.0001431852,0.000008824586,0.00006451961,0.00008751175,0.000008732681,0.00054825173],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99141407,0.004314546,0.00058627425,0.00079991634,0.0020147692,0.0008704363],"domain_scores_gemma":[0.8854943,0.06875053,0.020179577,0.0027631135,0.013030615,0.009781898],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007805523,0.00030166196,0.00041571129,0.005157612,0.0028630118,0.0029269713,0.0013957402,0.0010472479,0.0020262415],"category_scores_gemma":[0.061599486,0.00050290034,0.00016023291,0.0045376197,0.0021331364,0.0044298917,0.002997827,0.0015478492,0.0003479936],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00014358315,0.0013157371,0.669469,0.0002171801,0.00002702413,0.0009925697,0.29476866,0.0000627695,0.0012321023,0.0006285951,0.0010412364,0.030101495],"study_design_scores_gemma":[0.000031408184,0.00067154365,0.3825352,0.00015250104,0.00001904247,0.0012177125,0.6085589,0.0008065532,0.0005769348,0.00038440136,0.0050013233,0.000044497472],"about_ca_topic_score_codex":0.0051866216,"about_ca_topic_score_gemma":0.010546165,"teacher_disagreement_score":0.007805523,"about_ca_system_score_codex":0.0009886025,"about_ca_system_score_gemma":0.0020211258,"threshold_uncertainty_score":0.04128009},"labels":[],"label_agreement":null},{"id":"W1983866039","doi":"10.5555/2820518.2820590","title":"A dataset of the activity of the git super-repository of Linux in 2012","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Queen's University; Polytechnique Montréal; University of Victoria","funders":"","keywords":"Linux kernel; Computer science; World Wide Web; Kernel (algebra); Process (computing); Operating system","score_opus":0.026511254488833996,"score_gpt":0.26521439057043333,"score_spread":0.23870313608159935,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1983866039","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.049618937,0.000546956,0.0009699727,0.00029250997,0.00009592134,0.00008484401,0.9432911,0.0016572207,0.0034425575],"genre_scores_gemma":[0.017483931,0.000125683,0.0017652198,0.000045566187,0.000021947753,0.00009727862,0.979192,0.00009285516,0.0011755425],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.9980057,0.00019600283,0.0003083722,0.0004605524,0.000765857,0.00026356004],"domain_scores_gemma":[0.9934098,0.0010693164,0.0013049253,0.001230112,0.0022089235,0.00077683583],"candidate_categories":["bibliometrics"],"consensus_categories":[],"category_scores_codex":[0.0009349915,0.00096856867,0.0008114255,0.006771364,0.00077467423,0.0014939258,0.001253228,0.0011698089,0.0034434507],"category_scores_gemma":[0.0066146003,0.0003624724,0.0007937904,0.011183123,0.00035246985,0.001318617,0.0015700946,0.0011273565,0.007058319],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0008177292,0.00042828236,0.13062115,0.0018476354,0.00023235509,0.00065870956,0.0014851755,0.002888434,0.0041329674,0.0022037078,0.793766,0.060917865],"study_design_scores_gemma":[0.00013046498,0.00015975912,0.39463818,0.00034038708,0.0000975844,0.0008068338,0.0015384624,0.006532519,0.0037216947,0.0014398821,0.5904621,0.00013222171],"about_ca_topic_score_codex":0.029153619,"about_ca_topic_score_gemma":0.061781835,"teacher_disagreement_score":0.9932286,"about_ca_system_score_codex":0.0011758865,"about_ca_system_score_gemma":0.0015514282,"threshold_uncertainty_score":0.05796784},"labels":[],"label_agreement":null},{"id":"W1991705476","doi":"10.5555/2820518.2820577","title":"Mining StackOverflow to filter out off-topic IRC discussion","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Spam and Phishing Detection","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Computer science; Exploit; Filter (signal processing); Set (abstract data type); The Internet; World Wide Web; Online discussion; Stack (abstract data type); Information retrieval; Open source; Data science; Programming language; Computer security","score_opus":0.030339466299692388,"score_gpt":0.26051056375486775,"score_spread":0.23017109745517536,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1991705476","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8579358,0.003402831,0.10093258,0.001609072,0.00048542267,0.001007851,0.011615599,0.010386147,0.01262479],"genre_scores_gemma":[0.87172717,0.0009907193,0.07982119,0.00034353082,0.0005822,0.00084889465,0.03450322,0.000665212,0.010517868],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.996512,0.0007411117,0.0003211421,0.0008211769,0.00115789,0.00044673038],"domain_scores_gemma":[0.98206514,0.008163818,0.0033700971,0.0011613237,0.004396265,0.0008433696],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003991825,0.0018466505,0.0010352273,0.016247183,0.0017328694,0.0021353767,0.001532695,0.0020403913,0.0018518806],"category_scores_gemma":[0.017509656,0.00041737125,0.0010562466,0.0053073824,0.0006421164,0.0045924312,0.0016863927,0.0013101072,0.0027145608],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007740524,0.0013654395,0.29741895,0.001736716,0.00026624248,0.0017666571,0.0053595724,0.007168677,0.018968234,0.0043755947,0.059352197,0.60144764],"study_design_scores_gemma":[0.00015099577,0.00093026005,0.20250987,0.00076497684,0.0006147137,0.0021821214,0.008833152,0.59286815,0.06406649,0.018997299,0.10780596,0.00027603138],"about_ca_topic_score_codex":0.004492614,"about_ca_topic_score_gemma":0.0055245105,"teacher_disagreement_score":0.016247183,"about_ca_system_score_codex":0.0012569256,"about_ca_system_score_gemma":0.0023395836,"threshold_uncertainty_score":0.021111071},"labels":[],"label_agreement":null},{"id":"W1996488581","doi":"10.5555/2820518.2820567","title":"Do onboarding programs work","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Open Source Software Innovations","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Onboarding; Staffing; Computer science; Work (physics); Open source; Software engineering; Software bug; Software; Knowledge management; Data science; Engineering; Management","score_opus":0.04522810760757105,"score_gpt":0.2741533000331867,"score_spread":0.22892519242561565,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1996488581","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.8723241,0.0007912933,0.021557359,0.0062277084,0.00051951944,0.00028219502,0.00028614677,0.0022983167,0.09571322],"genre_scores_gemma":[0.9538592,0.0007378098,0.017367544,0.0012649463,0.00014924459,0.0002170024,0.000617747,0.0012912811,0.024495134],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.99452716,0.0021078174,0.00017426616,0.0011201731,0.0009922746,0.0010782868],"domain_scores_gemma":[0.96256506,0.014524729,0.00666405,0.0070027416,0.0036199396,0.005623567],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0062469663,0.0006311625,0.00042422168,0.002246559,0.0018063525,0.005152371,0.001469685,0.0011824038,0.015254752],"category_scores_gemma":[0.058488835,0.00046459775,0.00041673955,0.0021980687,0.0019959223,0.007923696,0.004102356,0.0013857662,0.004096181],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00037739787,0.00093778124,0.17371829,0.0004183625,0.00007639093,0.00032438492,0.024248222,0.0007482668,0.005054591,0.0211672,0.026995894,0.7459332],"study_design_scores_gemma":[0.00026822943,0.0017046131,0.37189227,0.0013091927,0.00034134035,0.001323368,0.043395337,0.008322862,0.011253849,0.054725897,0.50526404,0.00019889386],"about_ca_topic_score_codex":0.0038084975,"about_ca_topic_score_gemma":0.0041374443,"teacher_disagreement_score":0.015254752,"about_ca_system_score_codex":0.0015455202,"about_ca_system_score_gemma":0.0032738037,"threshold_uncertainty_score":0.051032305},"labels":[],"label_agreement":null},{"id":"W1998211793","doi":"10.5555/2820518.2820556","title":"What is the Gist?: understanding the use of public Gists on GitHub","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Open Source Software Innovations","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Computer science; GiST; Metadata; World Wide Web; Source code; Coding (social sciences); Variety (cybernetics); Information retrieval; Code (set theory); Medicine","score_opus":0.20485367274747782,"score_gpt":0.30147486018912495,"score_spread":0.09662118744164713,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W1998211793","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9875029,0.00041955916,0.0036520846,0.0014059538,0.0000107346,0.000035229972,0.00048646514,0.00018535941,0.0063017327],"genre_scores_gemma":[0.9918412,0.00059042504,0.0046255304,0.00026977854,0.000020015694,0.00004406366,0.0007928151,0.00023515832,0.001581079],"study_design_codex":"qualitative","study_design_gemma":"observational","domain_scores_codex":[0.9978362,0.0009364442,0.00016647745,0.00032812086,0.00050968246,0.00022309461],"domain_scores_gemma":[0.9813074,0.011838903,0.003234493,0.00092636375,0.0020943068,0.00059857307],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0028441902,0.00032068542,0.00027503123,0.0063198754,0.0012374016,0.0046378267,0.00072244107,0.0008561657,0.001550613],"category_scores_gemma":[0.02325305,0.00036661487,0.00031507472,0.0057710516,0.002354075,0.013163895,0.0025903163,0.00081947044,0.00076540967],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00016842438,0.00008918274,0.27012214,0.00044563398,0.000040508447,0.0011827392,0.6301693,0.00023266848,0.0034901323,0.004953411,0.006065826,0.08304002],"study_design_scores_gemma":[0.000009872341,0.00009213121,0.42883945,0.0004920072,0.00006359384,0.0013939824,0.51049614,0.0055355164,0.0020686658,0.0052131587,0.045667343,0.00012814668],"about_ca_topic_score_codex":0.017534627,"about_ca_topic_score_gemma":0.025606759,"teacher_disagreement_score":0.017534627,"about_ca_system_score_codex":0.0018705496,"about_ca_system_score_gemma":0.0010400614,"threshold_uncertainty_score":0.03486508},"labels":[],"label_agreement":null},{"id":"W2041589121","doi":"10.5555/2820518.2820597","title":"The Firefox temporal defect dataset","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Computer science; Software bug; Process (computing); Software; Plan (archaeology); Data mining; Data science; Geography","score_opus":0.030207281263329028,"score_gpt":0.28240052547346595,"score_spread":0.2521932442101369,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2041589121","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.07422754,0.0013300122,0.0021242949,0.00051778374,0.00012702253,0.0002400476,0.9140127,0.0036336228,0.0037870803],"genre_scores_gemma":[0.020450784,0.0002843566,0.005572837,0.00011885214,0.000037306014,0.00020261767,0.9718784,0.000115445866,0.0013393163],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9986951,0.000120478006,0.00016758201,0.00036249877,0.0004977438,0.00015652094],"domain_scores_gemma":[0.99709225,0.0007083876,0.0005041599,0.0005778714,0.0008228387,0.00029435626],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0012245564,0.0015040167,0.0006770855,0.0070736483,0.0007636699,0.00096230925,0.0020002269,0.0015365207,0.002736955],"category_scores_gemma":[0.0048390585,0.00041899376,0.0010749103,0.0048286514,0.00038262352,0.0012046547,0.00099546,0.0010335515,0.0039864425],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00079582236,0.0009345521,0.08896516,0.0019552219,0.0003254124,0.0012932018,0.0004995013,0.005266598,0.005905402,0.0021573105,0.7999632,0.091938674],"study_design_scores_gemma":[0.00074469694,0.00069510005,0.2622463,0.0004957426,0.0002899564,0.0032359173,0.00076869014,0.025987111,0.0081309965,0.0033234123,0.6938828,0.00019928733],"about_ca_topic_score_codex":0.027174866,"about_ca_topic_score_gemma":0.051422946,"teacher_disagreement_score":0.027174866,"about_ca_system_score_codex":0.0014230446,"about_ca_system_score_gemma":0.0014793462,"threshold_uncertainty_score":0.0540334},"labels":[],"label_agreement":null},{"id":"W2064653370","doi":"10.5555/2820518.2820527","title":"Co-evolution of infrastructure and source code: an empirical study","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":37,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Computer science; Server; Source code; Empirical research; Critical infrastructure; Database; Process (computing); Software engineering; Operating system; Programming language; Computer security","score_opus":0.02434610871695176,"score_gpt":0.31298362237117433,"score_spread":0.2886375136542226,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2064653370","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9982717,0.00017150806,0.0004851232,0.00011869745,0.000003946512,0.000047010693,0.00012835872,0.000012638669,0.0007609932],"genre_scores_gemma":[0.9988624,0.00008042626,0.0005007181,0.000032105017,0.0000069718217,0.000046794125,0.0002439431,0.000014957356,0.00021163041],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.977428,0.011702174,0.0018929503,0.0029060922,0.0047408114,0.0013299136],"domain_scores_gemma":[0.5008617,0.37992585,0.070991464,0.019274658,0.021189632,0.007756779],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.019101763,0.000398162,0.00054407504,0.0052620205,0.0012973772,0.0027510345,0.0020821462,0.0016544028,0.0030230195],"category_scores_gemma":[0.15376776,0.00070226274,0.00065541675,0.0071286913,0.0032284479,0.005493152,0.0032011957,0.002946265,0.00079474435],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000060964587,0.00029696737,0.99325174,0.000030128296,0.000056557077,0.00014016232,0.002464255,0.00008750714,0.00007855341,0.00014069478,0.00012298881,0.0032694228],"study_design_scores_gemma":[0.000010960483,0.00027444493,0.9880415,0.00004826335,0.000060111128,0.00060220866,0.007756908,0.0019837827,0.00019214714,0.00016471493,0.0008479757,0.00001695205],"about_ca_topic_score_codex":0.006286184,"about_ca_topic_score_gemma":0.006984899,"teacher_disagreement_score":0.019101763,"about_ca_system_score_codex":0.0014785766,"about_ca_system_score_gemma":0.0016719532,"threshold_uncertainty_score":0.10102099},"labels":[],"label_agreement":null},{"id":"W2067250880","doi":"10.5555/2820518.2820569","title":"Organizational volatility and post-release defects: a replication case study using data from Google Chrome","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Outsourcing; Popularity; Volatility (finance); Business; Computer science; Directory; Replicate; World Wide Web; Knowledge management; Marketing; Finance; Operating system","score_opus":0.060518256411597054,"score_gpt":0.3147072335686175,"score_spread":0.2541889771570205,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2067250880","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9983199,0.000065348766,0.0006917576,0.000066235894,0.000004828394,0.000081809994,0.0004628222,0.000027292828,0.00027995848],"genre_scores_gemma":[0.9963921,0.000047968522,0.0019701854,0.00003771475,0.000013839199,0.00012336436,0.0011033025,0.000021939526,0.00028945308],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9909151,0.0047658063,0.00082910585,0.0014349901,0.0015451835,0.00050975324],"domain_scores_gemma":[0.90250343,0.054310214,0.015059375,0.015594817,0.010884887,0.0016472713],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.010282141,0.0005558202,0.0006591169,0.00355258,0.0012536842,0.0015292554,0.002075019,0.0017000554,0.0008709373],"category_scores_gemma":[0.051926654,0.0004851667,0.0011922937,0.0043948977,0.0013041538,0.001680185,0.0013786611,0.0013928548,0.0004046446],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00058211817,0.0015956595,0.9559567,0.0002614817,0.0003656178,0.003179928,0.013209667,0.002551248,0.0020728717,0.00052739395,0.0015693677,0.018127955],"study_design_scores_gemma":[0.00019154155,0.0014426775,0.95891935,0.000100358826,0.00027408605,0.0018875883,0.017628497,0.011760616,0.003226514,0.0006476085,0.0037767484,0.00014441309],"about_ca_topic_score_codex":0.026591904,"about_ca_topic_score_gemma":0.026666792,"teacher_disagreement_score":0.98971784,"about_ca_system_score_codex":0.0014704914,"about_ca_system_score_gemma":0.0011633792,"threshold_uncertainty_score":0.054377854},"labels":[],"label_agreement":null},{"id":"W2074549053","doi":"10.5555/2820518.2820578","title":"An insight into the unresolved questions at stack overflow","year":2015,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Expert finding and Q&A systems","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Stack (abstract data type); Computer science; Popularity; Recall; Data science; Exploratory research; Information retrieval; Theoretical computer science; Artificial intelligence; Cognitive psychology; Programming language; Psychology; Social psychology","score_opus":0.025449479710167613,"score_gpt":0.2739315046489646,"score_spread":0.24848202493879698,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2074549053","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.95660096,0.0011716456,0.030357962,0.0032576344,0.000057189594,0.00012663538,0.0014025235,0.0005171201,0.006508388],"genre_scores_gemma":[0.9870438,0.00027361797,0.009705952,0.0003257081,0.00005449912,0.00003531132,0.0006483879,0.00006127034,0.0018514978],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9967422,0.0013428609,0.00032132724,0.0004597448,0.0008247335,0.00030914563],"domain_scores_gemma":[0.9462408,0.040810723,0.0058456305,0.0017767784,0.004364764,0.0009613182],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005748033,0.00044428595,0.0003981444,0.004280173,0.0014431792,0.002302072,0.0008261624,0.0016792319,0.003299846],"category_scores_gemma":[0.04334348,0.0003264532,0.0004965007,0.0027053594,0.001204041,0.006007239,0.0015522697,0.0013253329,0.00063698477],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00089154334,0.0005123828,0.6785446,0.0009431565,0.00011393322,0.0035544257,0.06436989,0.0051441514,0.024373697,0.026261775,0.008603002,0.18668747],"study_design_scores_gemma":[0.000067898785,0.0008097932,0.7280408,0.0006004469,0.00018727961,0.0056708897,0.04268225,0.07659919,0.018192276,0.067244545,0.05957614,0.00032850544],"about_ca_topic_score_codex":0.0032283647,"about_ca_topic_score_gemma":0.003829961,"teacher_disagreement_score":0.005748033,"about_ca_system_score_codex":0.00086450565,"about_ca_system_score_gemma":0.0008442328,"threshold_uncertainty_score":0.030398846},"labels":[],"label_agreement":null},{"id":"W2079972013","doi":"10.5555/2664446.2664462","title":"Mining challenge 2012: the Android platform","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":28,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Queen's University","funders":"","keywords":"Android (operating system); Computer science; Data science; World Wide Web; Android application; Software; Android app; Operating system","score_opus":0.027878367104243282,"score_gpt":0.26460729060075805,"score_spread":0.23672892349651475,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2079972013","genre_codex":"dataset","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.29234445,0.022963798,0.062254928,0.12765819,0.015759427,0.0038936872,0.32046354,0.06511669,0.08954533],"genre_scores_gemma":[0.28544265,0.0053572264,0.089229226,0.008655172,0.003795067,0.0028692782,0.53364074,0.006912999,0.06409771],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.9833149,0.0034000212,0.0012161919,0.0023212868,0.008621025,0.0011265654],"domain_scores_gemma":[0.9651196,0.011080197,0.0019294337,0.007895457,0.010706387,0.0032689478],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012033563,0.0016545659,0.0018906152,0.0048384825,0.0030983281,0.005481907,0.0031493232,0.0042482405,0.0050166473],"category_scores_gemma":[0.043675374,0.0009332369,0.0016000643,0.0046825847,0.0012570615,0.0060783573,0.006122851,0.0042640655,0.007082395],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00059351855,0.00041079224,0.008381545,0.00088787073,0.00013144367,0.0005510005,0.00078151777,0.001460745,0.00232036,0.005187357,0.8924989,0.08679497],"study_design_scores_gemma":[0.00043071044,0.0005357007,0.039245863,0.00042478347,0.00011548254,0.0012236774,0.0016170483,0.028261837,0.008735622,0.009341521,0.9098697,0.00019816846],"about_ca_topic_score_codex":0.020626204,"about_ca_topic_score_gemma":0.02885329,"teacher_disagreement_score":0.020626204,"about_ca_system_score_codex":0.0018791122,"about_ca_system_score_gemma":0.0053015705,"threshold_uncertainty_score":0.063640356},"labels":[],"label_agreement":null},{"id":"W2081647265","doi":"10.5555/2664446.2664469","title":"The evolution of the social programmer","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Open Source Software Innovations","field":"Computer Science","cited_by":9,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Victoria","funders":"","keywords":"Insourcing; Outsourcing; Crowdsourcing; Crowdsourcing software development; Programmer; Software development; Software; Knowledge management; Computer science; World Wide Web; Social software engineering; Social media; Social software; Social computing; Software engineering; Data science; Business; Software development process; Software construction; Marketing","score_opus":0.013594462858546503,"score_gpt":0.26177028209367786,"score_spread":0.24817581923513135,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2081647265","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.11863817,0.053665332,0.15577649,0.13678902,0.003501168,0.00041123034,0.0015881454,0.00088530185,0.5287451],"genre_scores_gemma":[0.8798707,0.018787863,0.045012504,0.008823162,0.0027252415,0.00036350777,0.0004833253,0.00028392277,0.043649793],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99522793,0.0021644426,0.00013255782,0.0011903686,0.0010452018,0.00023948247],"domain_scores_gemma":[0.9839835,0.008857934,0.0015973552,0.0022182243,0.0017318652,0.0016111613],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0037220465,0.000547964,0.0004954446,0.0028470966,0.004032792,0.008000784,0.0013868364,0.0038882904,0.011137889],"category_scores_gemma":[0.013133324,0.0004816532,0.00038080823,0.0024325103,0.015582945,0.011447814,0.0061959634,0.0032734445,0.0019360966],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000033187505,0.000036913127,0.0029617257,0.00022456057,0.000024010256,0.00017491779,0.0102413725,0.00042106272,0.00078819494,0.87024987,0.012642602,0.102201544],"study_design_scores_gemma":[0.000012518619,0.000053529526,0.00403371,0.00037845486,0.000015425827,0.00040308235,0.0038135587,0.0020598206,0.0004114537,0.32554278,0.66323054,0.000045163037],"about_ca_topic_score_codex":0.0041367332,"about_ca_topic_score_gemma":0.0037394257,"teacher_disagreement_score":0.011137889,"about_ca_system_score_codex":0.004015439,"about_ca_system_score_gemma":0.0026248943,"threshold_uncertainty_score":0.037259996},"labels":[],"label_agreement":null},{"id":"W2117757207","doi":"10.5555/2664446.2664448","title":"Towards improving bug tracking systems with game mechanisms","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":30,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Agile software development; Tracking system; Quality (philosophy); Tracking (education); Reputation; Software engineering; Computer security; Human–computer interaction; Artificial intelligence","score_opus":0.016023266705648975,"score_gpt":0.2417133949857941,"score_spread":0.22569012828014515,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2117757207","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.2284522,0.0013741672,0.74481153,0.0039394903,0.0001740057,0.0017247443,0.00018156979,0.014493053,0.004849255],"genre_scores_gemma":[0.41617188,0.0003601163,0.5800859,0.00043684753,0.00006574838,0.0006625094,0.00024971686,0.00042720762,0.0015401285],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9741469,0.014814288,0.0023566869,0.002788761,0.004832205,0.0010611378],"domain_scores_gemma":[0.8492488,0.07648979,0.020185366,0.027919468,0.021555187,0.0046014027],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.04116054,0.0018904903,0.0014686637,0.008433836,0.0017330033,0.0074062347,0.004907247,0.0023980138,0.002657325],"category_scores_gemma":[0.1769607,0.0015512309,0.0014540899,0.0035446703,0.0025819307,0.012927097,0.0061343783,0.0029240085,0.0011553917],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0007480231,0.0030292596,0.07757057,0.0018360469,0.00048863294,0.0003031157,0.0055301613,0.03334528,0.037492372,0.036396023,0.0077032195,0.79555726],"study_design_scores_gemma":[0.0009737936,0.004037373,0.03941377,0.0011048996,0.00087807275,0.0010390257,0.003641111,0.7403758,0.03698729,0.118830785,0.05214976,0.00056831114],"about_ca_topic_score_codex":0.002730855,"about_ca_topic_score_gemma":0.0038160179,"teacher_disagreement_score":0.04116054,"about_ca_system_score_codex":0.002179718,"about_ca_system_score_gemma":0.00435208,"threshold_uncertainty_score":0.21768034},"labels":[],"label_agreement":null},{"id":"W2131704937","doi":"10.5555/2664446.2664458","title":"Green mining: a methodology of relating software change to power consumption","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Green IT and Sustainability","field":"Engineering","cited_by":64,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Software; Computer science; Consumption (sociology); Power consumption; BitTorrent; Power (physics); Operating system; The Internet","score_opus":0.06653757338892163,"score_gpt":0.2953637917156099,"score_spread":0.22882621832668826,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2131704937","genre_codex":"methods","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":"methods","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.057325486,0.00024009263,0.9335084,0.00027191144,0.000046916557,0.0008674785,0.0017025825,0.0026287881,0.0034082772],"genre_scores_gemma":[0.28450263,0.00034286414,0.7087525,0.000103739156,0.0000554391,0.00169596,0.0017504556,0.0003656788,0.0024306693],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9925956,0.0019628038,0.0007990098,0.0018612833,0.0025641262,0.00021709118],"domain_scores_gemma":[0.97571975,0.013586905,0.004479558,0.003258305,0.0027252862,0.00023024609],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0056872685,0.0014501238,0.001344179,0.010401205,0.00070278597,0.0021577065,0.0018346547,0.0009977496,0.0020618995],"category_scores_gemma":[0.037587535,0.0008028345,0.0017732161,0.010126578,0.0010711913,0.0029106364,0.0016026474,0.0014029894,0.00073271245],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003234967,0.0009540045,0.17269623,0.0019334056,0.0010468845,0.00071403035,0.004296434,0.06841,0.02375638,0.026809,0.004882531,0.6941776],"study_design_scores_gemma":[0.00012096606,0.0014273193,0.17570774,0.0004927789,0.0005866411,0.0016991626,0.0029237631,0.659358,0.056129053,0.06528321,0.035853844,0.00041746447],"about_ca_topic_score_codex":0.0044399933,"about_ca_topic_score_gemma":0.005273161,"teacher_disagreement_score":0.010401205,"about_ca_system_score_codex":0.0011261163,"about_ca_system_score_gemma":0.001549939,"threshold_uncertainty_score":0.030077517},"labels":[],"label_agreement":null},{"id":"W2154813515","doi":"10.5555/2664446.2664465","title":"Do the stars align?: multidimensional analysis of Android's layered architecture","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Advanced Malware Detection Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"","keywords":"Android (operating system); Computer science; Popularity; Architecture; Locality; Operating system; Geography","score_opus":0.011744882525077764,"score_gpt":0.260679922151485,"score_spread":0.24893503962640723,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2154813515","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9033952,0.00053869915,0.09062338,0.0005401772,0.000023842533,0.000104342704,0.001086456,0.0007893522,0.0028985396],"genre_scores_gemma":[0.94329154,0.00019983939,0.05487932,0.000029334497,0.000015641812,0.000070428076,0.00083244615,0.000076236785,0.0006051478],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9987508,0.0003624834,0.00012427353,0.00020609805,0.00038992424,0.00016635505],"domain_scores_gemma":[0.9944271,0.0020433632,0.0013450355,0.0006800566,0.0011443818,0.0003601048],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014038279,0.0004789013,0.0004121613,0.0069761067,0.00082142185,0.0031201595,0.00048432918,0.00049892877,0.001110525],"category_scores_gemma":[0.01079828,0.00037422383,0.0009005535,0.0048596514,0.0007425351,0.0021530227,0.0020350805,0.0007129806,0.00043437124],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00032609954,0.00017930883,0.70332026,0.00030366873,0.00031259088,0.00078892504,0.009460963,0.019661425,0.014962285,0.017840967,0.003951376,0.22889206],"study_design_scores_gemma":[0.000020820036,0.00012902496,0.5744362,0.000110666544,0.00013185514,0.0009787832,0.008502584,0.3743278,0.0048131975,0.027103845,0.009310436,0.00013479007],"about_ca_topic_score_codex":0.0064370595,"about_ca_topic_score_gemma":0.005458452,"teacher_disagreement_score":0.0069761067,"about_ca_system_score_codex":0.00071322115,"about_ca_system_score_gemma":0.0006314453,"threshold_uncertainty_score":0.012799203},"labels":[],"label_agreement":null},{"id":"W2160841184","doi":"10.5555/2664446.2664451","title":"A linked data platform for mining software repositories","year":2012,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":41,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Concordia University","funders":"","keywords":"Computer science; Software versioning; Software; World Wide Web; Cloud computing; Software mining; BitTorrent tracker; Database; Software engineering; Software development; Data science; Software construction; Operating system","score_opus":0.05717894938690845,"score_gpt":0.3065874026592379,"score_spread":0.24940845327232944,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2160841184","genre_codex":"methods","genre_gemma":"software","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"software","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01223419,0.00092918915,0.5841338,0.0013581768,0.0002525233,0.0033408836,0.27210966,0.11557677,0.010064783],"genre_scores_gemma":[0.03464387,0.0007687688,0.47233188,0.00039523304,0.00009766586,0.0032980796,0.48077345,0.0038027407,0.0038882955],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99225956,0.001260577,0.0016189789,0.0015612367,0.0029942144,0.00030540256],"domain_scores_gemma":[0.98044854,0.0051992326,0.0024416954,0.0074548046,0.0032092175,0.0012464182],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0076530334,0.0016259234,0.0013581993,0.019198036,0.0025359676,0.005473047,0.003818718,0.0021559289,0.0073091793],"category_scores_gemma":[0.031804666,0.0015944192,0.0023374956,0.021194346,0.0007974037,0.00890964,0.008550226,0.0029873874,0.006393253],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0014404076,0.0010505414,0.029736903,0.0048840973,0.0014132732,0.0016804414,0.002915089,0.032135494,0.0148472795,0.11898215,0.30751058,0.48340377],"study_design_scores_gemma":[0.0005185564,0.0003388406,0.018770913,0.0009799444,0.00036259153,0.00088438694,0.0009625295,0.13869174,0.019515572,0.14143911,0.67706376,0.00047210458],"about_ca_topic_score_codex":0.009962943,"about_ca_topic_score_gemma":0.010423034,"teacher_disagreement_score":0.019198036,"about_ca_system_score_codex":0.002054416,"about_ca_system_score_gemma":0.0049691745,"threshold_uncertainty_score":0.04047358},"labels":[],"label_agreement":null},{"id":"W2907438654","doi":"","title":"Revisiting \"Programmers' Build Errors\" in the Visual Studio Context","year":2018,"lang":"en","type":"article","venue":"Mining Software Repositories","topic":"Software Engineering Research","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McGill University","funders":"","keywords":"Deliverable; Computer science; TRACE (psycholinguistics); Context (archaeology); Replicate; Workspace; Studio; Codebase; Code (set theory); Software engineering; Human–computer interaction; Programming language; Source code; World Wide Web; Artificial intelligence; Systems engineering","score_opus":0.01943302576174965,"score_gpt":0.3028428703419978,"score_spread":0.28340984458024815,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2907438654","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.98104846,0.0013913554,0.0101268925,0.001208826,0.00007424561,0.00007540589,0.0011333587,0.0003172558,0.0046241465],"genre_scores_gemma":[0.9875158,0.0004568835,0.008582102,0.0003561976,0.000086058746,0.0000687632,0.0016066785,0.00026204082,0.001065485],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9809806,0.006800051,0.0019134311,0.0034489897,0.0056789354,0.0011780631],"domain_scores_gemma":[0.8261028,0.100335665,0.041094806,0.015728327,0.014734277,0.002004064],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.013781794,0.0006318505,0.0005083072,0.009184364,0.0014392821,0.0039651375,0.0013300172,0.0009808649,0.0010321592],"category_scores_gemma":[0.11348235,0.00049982965,0.00060740515,0.009544346,0.001914285,0.0060046,0.0038991577,0.0019211192,0.000394993],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001772377,0.00017115835,0.87633836,0.00052997086,0.00013585619,0.0007265903,0.0275237,0.0009882682,0.0015180247,0.0025084605,0.0053024734,0.08407994],"study_design_scores_gemma":[0.00002782595,0.00021838475,0.9237546,0.00053379586,0.00016534458,0.001436999,0.032774717,0.007738865,0.0025405379,0.006870514,0.02386049,0.00007794557],"about_ca_topic_score_codex":0.010240169,"about_ca_topic_score_gemma":0.024988089,"teacher_disagreement_score":0.013781794,"about_ca_system_score_codex":0.0010742162,"about_ca_system_score_gemma":0.001840143,"threshold_uncertainty_score":0.07288599},"labels":[],"label_agreement":null}]}