{"meta":{"query_hash":"39236c7140f9","filters":{"venue":"Balisage series on markup technologies"},"cohort_total":30,"direct_labels_cover":1,"predictions_cover":30,"exported":30,"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/39236c7140f9","api":"https://metacan.xera.ac/api/v1/cohort?venue=Balisage+series+on+markup+technologies"},"results":[{"id":"W2183772626","doi":"10.4242/balisagevol8.huitfeldt02","title":"The MLCD Overlap Corpus (MOC)","year":2012,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Natural language processing; Artificial intelligence","score_opus":0.033983899520821315,"score_gpt":0.21915949484324737,"score_spread":0.18517559532242606,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2183772626","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":[0.06099697,0.004245605,0.23153883,0.004842727,0.0019353505,0.0047187223,0.39020216,0.017753148,0.2837666],"genre_scores_gemma":[0.15238655,0.0012935952,0.30263135,0.0011464537,0.00063871616,0.009468487,0.47274023,0.008366085,0.051328626],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99408156,0.0023294585,0.0006982656,0.0010194655,0.0016154511,0.0002558956],"domain_scores_gemma":[0.9835789,0.0073838443,0.0007922411,0.0035939082,0.00397207,0.00067909335],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004567735,0.0007500019,0.00086818327,0.009515409,0.003908332,0.0045240917,0.0020647794,0.0016014682,0.052113455],"category_scores_gemma":[0.025614394,0.0008070709,0.00041945215,0.010768362,0.0022636103,0.004921978,0.007354988,0.0024431066,0.018961761],"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":[0.00056423154,0.0002076244,0.0027082209,0.0034035302,0.00004262263,0.0007313441,0.013012974,0.001390664,0.009466377,0.09774813,0.5567619,0.31396243],"study_design_scores_gemma":[0.00007293191,0.000028970631,0.002889178,0.00043197453,0.00001589526,0.0004084916,0.0015564371,0.0017810694,0.004010731,0.009298436,0.9794455,0.000060465038],"about_ca_topic_score_codex":0.007998256,"about_ca_topic_score_gemma":0.009506025,"teacher_disagreement_score":0.052113455,"about_ca_system_score_codex":0.0021636228,"about_ca_system_score_gemma":0.0044896565,"threshold_uncertainty_score":0.17433691},"labels":[],"label_agreement":null},{"id":"W2273502271","doi":"10.4242/balisagevol8.jordanous01","title":"Contemporary transformation of ancient documents for recording and retrieving maximum information: when one form of markup is not enough","year":2012,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":6,"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 Prince Edward Island","funders":"","keywords":"Markup language; Arabic; Computer science; Annotation; RDF; Transformation (genetics); Publication; Information retrieval; World Wide Web; Natural language processing; Linguistics; Artificial intelligence; XML; Semantic Web","score_opus":0.03342159134163366,"score_gpt":0.2468725924931488,"score_spread":0.21345100115151514,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2273502271","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":[0.108073905,0.009383893,0.49029648,0.03982624,0.003378823,0.00037415302,0.0014073893,0.0028155798,0.34444356],"genre_scores_gemma":[0.55192184,0.0072468324,0.386392,0.0025201242,0.00061842526,0.00033046087,0.0015620046,0.0025347357,0.04687366],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9923578,0.003109853,0.000591511,0.000815914,0.0028945154,0.00023040066],"domain_scores_gemma":[0.97401005,0.00937868,0.0012660965,0.010745821,0.004222141,0.00037721125],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012610856,0.00044970796,0.00044860225,0.0035720277,0.0040314435,0.0077763447,0.0014291396,0.002148206,0.0057496005],"category_scores_gemma":[0.034771655,0.00039183162,0.00032843737,0.0059979786,0.011094461,0.019366726,0.0045318077,0.0035352933,0.0025636577],"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":[0.00012060198,0.0000658025,0.0015681932,0.0008288098,0.000024219073,0.0005771019,0.07192428,0.0012551929,0.008359072,0.6033808,0.016800944,0.29509503],"study_design_scores_gemma":[0.000016645457,0.000038983897,0.0012150526,0.0009636382,0.000030281084,0.0010153259,0.021142056,0.0018346186,0.013700549,0.14045873,0.8194973,0.0000868231],"about_ca_topic_score_codex":0.0028528152,"about_ca_topic_score_gemma":0.002907866,"teacher_disagreement_score":0.012610856,"about_ca_system_score_codex":0.002916761,"about_ca_system_score_gemma":0.004149251,"threshold_uncertainty_score":0.066693366},"labels":[],"label_agreement":null},{"id":"W2277519993","doi":"10.4242/balisagevol5.briquet01","title":"A Virtualization-Based Retrieval and Update API for XML-Encoded Corpora","year":2010,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University","funders":"","keywords":"Computer science; XML; Information retrieval; Markup language; World Wide Web; Simple API for XML; Well-formed document; XML framework; SGML; Representation (politics); Virtualization; Streaming XML; Document Structure Description; Document type definition; XML Signature; Operating system; Cloud computing","score_opus":0.009566712296790668,"score_gpt":0.2309762155939664,"score_spread":0.2214095032971757,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2277519993","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":[0.0057415687,0.00025308295,0.70551187,0.00027944354,0.00012852912,0.0003902909,0.0047829296,0.2746464,0.008265859],"genre_scores_gemma":[0.16066097,0.0008547547,0.7272001,0.00073313585,0.00021210863,0.001713193,0.03341822,0.045130145,0.030077396],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99861145,0.0003076115,0.00022598573,0.0002516513,0.0005065493,0.00009676436],"domain_scores_gemma":[0.9957467,0.0016176185,0.00029403082,0.0016472734,0.00052096345,0.00017349292],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027437457,0.001051721,0.0010420558,0.0023039104,0.00084623747,0.003528894,0.0025795985,0.00096322974,0.01485512],"category_scores_gemma":[0.009531323,0.0009360973,0.00093447947,0.0025120813,0.0008107592,0.0053865598,0.003555312,0.0017139594,0.009161875],"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":[0.001777708,0.0003399208,0.0031225197,0.0010783321,0.00025594173,0.0014760299,0.00212136,0.009053993,0.066690184,0.073901586,0.1843223,0.6558601],"study_design_scores_gemma":[0.0004240326,0.0002602451,0.004382635,0.00025996295,0.00017997246,0.0026643167,0.0005151586,0.25446025,0.14031155,0.065874085,0.53018874,0.00047904605],"about_ca_topic_score_codex":0.0024990041,"about_ca_topic_score_gemma":0.0026243676,"teacher_disagreement_score":0.01485512,"about_ca_system_score_codex":0.001263306,"about_ca_system_score_gemma":0.0011541574,"threshold_uncertainty_score":0.049695313},"labels":[],"label_agreement":null},{"id":"W2397784846","doi":"10.4242/balisagevol15.gotti01","title":"XSDGuide - Automated Generation of Web Interfaces from XML Schemas: A Case Study for Suspicious Activity Reporting","year":2015,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Validator; XML Schema (W3C); XML; XML validation; Document Structure Description; World Wide Web; Streaming XML; XML Schema Editor; Schema (genetic algorithms); Efficient XML Interchange; JavaScript; XML framework; User interface; Information retrieval; Document type definition; Programming language","score_opus":0.13638877290988888,"score_gpt":0.34546901583523076,"score_spread":0.20908024292534189,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2397784846","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":[0.56253046,0.0007069411,0.38708875,0.0015062515,0.00013914716,0.0017907994,0.0014490654,0.030759262,0.014029412],"genre_scores_gemma":[0.48348346,0.00046145625,0.5004401,0.0003200155,0.000022233125,0.00049705367,0.0023491608,0.0032192483,0.009207193],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9953961,0.0025435807,0.0003399919,0.00043752498,0.0011222719,0.00016049338],"domain_scores_gemma":[0.98652565,0.009266423,0.00046060834,0.0023496235,0.0011108461,0.00028677168],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006331678,0.0006453641,0.00052775175,0.0010794342,0.000912309,0.0022830649,0.0025112557,0.00217196,0.0020328832],"category_scores_gemma":[0.018942093,0.0006823409,0.000788246,0.00096095237,0.0013887528,0.003646077,0.0019496917,0.0016962852,0.0010364777],"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":[0.0021650624,0.005177516,0.036173288,0.0031021636,0.00024018617,0.03317592,0.061543282,0.055940297,0.10379259,0.071009964,0.03990439,0.5877754],"study_design_scores_gemma":[0.00085859676,0.0018924986,0.012488865,0.0007215644,0.00020582027,0.016609773,0.009926471,0.32080835,0.28169292,0.013871811,0.34053025,0.000393068],"about_ca_topic_score_codex":0.0035932881,"about_ca_topic_score_gemma":0.0036439074,"teacher_disagreement_score":0.006331678,"about_ca_system_score_codex":0.0009884372,"about_ca_system_score_gemma":0.0010325472,"threshold_uncertainty_score":0.03348553},"labels":[],"label_agreement":null},{"id":"W2405269923","doi":"10.4242/balisagevol1.muldner01","title":"Secure Publishing using Schema-level Role-based Access Control Policies for Fragments of XML Documents","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Access Control and Trust","field":"Social Sciences","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":"Acadia University","funders":"","keywords":"Computer science; Document Structure Description; XML Encryption; XML Schema Editor; XML; Schema (genetic algorithms); Access control; Efficient XML Interchange; Parsing; Encryption; XML validation; Information retrieval; Database; Computer security; World Wide Web; Programming language","score_opus":0.05453315838117753,"score_gpt":0.3539137275040181,"score_spread":0.2993805691228406,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2405269923","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":[0.014155795,0.0004637283,0.97612983,0.0005857342,0.00009773052,0.0003103101,0.00010318994,0.0028136808,0.005339861],"genre_scores_gemma":[0.2983551,0.001739259,0.68746936,0.0003806512,0.00024145389,0.00047869896,0.00074139476,0.0006219139,0.009972107],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.98843426,0.0043456336,0.0018569763,0.0011112781,0.0036282684,0.00062360003],"domain_scores_gemma":[0.987021,0.0033644137,0.0009447806,0.0072152237,0.0010687654,0.00038571443],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009914662,0.0005646513,0.0008246829,0.0012257402,0.0010736013,0.007231507,0.0022480944,0.0018696968,0.0020081114],"category_scores_gemma":[0.013677187,0.0008731651,0.0016291157,0.0014878141,0.002690602,0.0077727516,0.0029919683,0.0026341006,0.0021100764],"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":[0.00041643876,0.00035838518,0.002080091,0.00038530256,0.000120193945,0.00089096936,0.0018496189,0.034485273,0.046260986,0.74431753,0.0057973084,0.16303797],"study_design_scores_gemma":[0.0002554585,0.0004743068,0.0010515304,0.00027252454,0.00021821183,0.0022616233,0.0005073529,0.4113643,0.15697385,0.23798119,0.18846384,0.00017580159],"about_ca_topic_score_codex":0.0015294069,"about_ca_topic_score_gemma":0.0009949347,"teacher_disagreement_score":0.009914662,"about_ca_system_score_codex":0.0017048264,"about_ca_system_score_gemma":0.0028714193,"threshold_uncertainty_score":0.052434385},"labels":[],"label_agreement":null},{"id":"W2406907785","doi":"10.4242/balisagevol4.cameron01","title":"Parallel Bit Stream Technology as a Foundation for XML Parsing Performance","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Advanced Database Systems and Queries","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Computer science; Parsing; SIMD; Byte; Compiler; XML; Bitstream; Automatic summarization; Programming language; Simple API for XML; Computer architecture; Stream processing; Parallel computing; Operating system; Efficient XML Interchange; Artificial intelligence; XML Signature","score_opus":0.013928012532668961,"score_gpt":0.25957456622759684,"score_spread":0.24564655369492788,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2406907785","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":[0.03247726,0.009916172,0.8894401,0.0023941302,0.00073701836,0.000235877,0.00074998813,0.00886827,0.0551811],"genre_scores_gemma":[0.32525504,0.014707964,0.63959664,0.00083219475,0.00050755515,0.0004032494,0.0016972651,0.001816779,0.015183371],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9990182,0.00013639117,0.00011560167,0.00013436233,0.00053085666,0.0000645255],"domain_scores_gemma":[0.9974948,0.0011059718,0.000121744575,0.00040309998,0.00082470203,0.000049781047],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0016335734,0.0005809961,0.0005983858,0.0010346264,0.00046876218,0.0029930025,0.0011846963,0.00076874555,0.0088625755],"category_scores_gemma":[0.006073335,0.0005562291,0.00039401127,0.001907339,0.0007227131,0.005178384,0.0007820007,0.0016850693,0.0036558926],"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":[0.0007394789,0.0001304403,0.0015342669,0.0012236385,0.00005321451,0.00051500596,0.00041728315,0.02952498,0.19273576,0.3345457,0.016773159,0.42180702],"study_design_scores_gemma":[0.00008469303,0.00045843422,0.0012650613,0.00038087388,0.00013694524,0.0013369008,0.00022931167,0.26365742,0.37938023,0.1893987,0.16352223,0.00014926151],"about_ca_topic_score_codex":0.00093655364,"about_ca_topic_score_gemma":0.0006477581,"teacher_disagreement_score":0.0088625755,"about_ca_system_score_codex":0.00092787907,"about_ca_system_score_gemma":0.0011195857,"threshold_uncertainty_score":0.029648304},"labels":[],"label_agreement":null},{"id":"W2406990242","doi":"10.4242/balisagevol3.muldner01","title":"XSAQCT: XML Queryable Compressor","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Advanced Database Systems and Queries","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":"Acadia University","funders":"","keywords":"Computer science; XML; Database; Gas compressor; Accounting; Software engineering; World Wide Web; Engineering; Business; Mechanical engineering","score_opus":0.009228316138441564,"score_gpt":0.22979839607798255,"score_spread":0.22057007993954098,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2406990242","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":[0.031259574,0.0019971973,0.69327605,0.0012068137,0.0005075391,0.0012552085,0.017994283,0.23521854,0.017284814],"genre_scores_gemma":[0.3593138,0.0031442903,0.50900364,0.0023055435,0.0005295068,0.0016054433,0.052034676,0.017662395,0.05440074],"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99814653,0.0001641092,0.00019486403,0.00019683263,0.0011531496,0.00014454684],"domain_scores_gemma":[0.9977946,0.0006887353,0.00014430007,0.00057025533,0.00067468546,0.00012740547],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013464045,0.0012939353,0.00078098266,0.0018425041,0.0005806288,0.001678065,0.0024655939,0.0011646793,0.020469157],"category_scores_gemma":[0.005213108,0.00040439877,0.0005735611,0.0025182732,0.0011524189,0.003200038,0.0018687849,0.0012675414,0.0038616904],"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":[0.0043262118,0.00024823225,0.0033417142,0.001983983,0.00018595974,0.0012284304,0.0009822974,0.012631389,0.15668939,0.04886974,0.23739117,0.53212154],"study_design_scores_gemma":[0.00074683956,0.00080910255,0.0023704632,0.0001993226,0.00013804436,0.001924884,0.0004163267,0.18792345,0.32749635,0.019424865,0.45827073,0.00027963487],"about_ca_topic_score_codex":0.005795692,"about_ca_topic_score_gemma":0.003264559,"teacher_disagreement_score":0.020469157,"about_ca_system_score_codex":0.00094902894,"about_ca_system_score_gemma":0.0010862541,"threshold_uncertainty_score":0.06847614},"labels":[],"label_agreement":null},{"id":"W2413440334","doi":"10.4242/balisagevol7.mercier01","title":"Including XSLT stylesheets testing in continuous integration process","year":2011,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":16,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"XSLT; Code refactoring; Computer science; Programming language; Java; Process (computing); Software engineering; XML; Software; Operating system","score_opus":0.12505217525480483,"score_gpt":0.25729339954486585,"score_spread":0.13224122429006102,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2413440334","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":[0.019994114,0.00020132042,0.9620335,0.00066130934,0.00012027115,0.00027731367,0.00006353966,0.011124488,0.0055240714],"genre_scores_gemma":[0.21538077,0.00026781711,0.77561307,0.0005122908,0.00010582522,0.00023909066,0.00029726184,0.001727579,0.005856256],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98545945,0.0051250537,0.00095602707,0.0015686163,0.006327996,0.00056292774],"domain_scores_gemma":[0.965217,0.015751874,0.0020522,0.010274777,0.0058242176,0.0008799566],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012373243,0.0010917744,0.0007474027,0.0016898352,0.00079308904,0.003269315,0.002326713,0.0020006667,0.0045606615],"category_scores_gemma":[0.034904525,0.0010356017,0.00082398107,0.00071351853,0.0018989877,0.006518419,0.003015448,0.00225171,0.002272774],"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":[0.0007737773,0.0008350697,0.009839624,0.0005419385,0.0000803459,0.002143012,0.0038933281,0.024941456,0.07856748,0.037493244,0.0071080974,0.8337827],"study_design_scores_gemma":[0.00065576227,0.0029875194,0.007436081,0.0015115975,0.0002570599,0.0047529954,0.0011295923,0.34471345,0.29327664,0.124446355,0.21828595,0.0005470176],"about_ca_topic_score_codex":0.00089881587,"about_ca_topic_score_gemma":0.00088040193,"teacher_disagreement_score":0.012373243,"about_ca_system_score_codex":0.0006408197,"about_ca_system_score_gemma":0.0019076822,"threshold_uncertainty_score":0.06543672},"labels":[],"label_agreement":null},{"id":"W2414491991","doi":"10.4242/balisagevol3.sperberg-mcqueen01","title":"Formal and informal meaning from documents through skeleton sentences","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; Meaning (existential); Sentence; Set (abstract data type); Natural language; Natural (archaeology); Semantics (computer science); Linguistics; Natural language processing; Complement (music); Natural language generation; Artificial intelligence; Epistemology; Programming language; Philosophy; History","score_opus":0.009119940111175638,"score_gpt":0.2504557326011053,"score_spread":0.24133579248992967,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2414491991","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":[0.014082254,0.0027158353,0.93002254,0.0040143956,0.00046848672,0.0003418028,0.0016054595,0.0010049492,0.045744207],"genre_scores_gemma":[0.2554068,0.0040929527,0.7151125,0.0015376643,0.00068858213,0.0008706337,0.0046454393,0.0008809838,0.016764509],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9956044,0.0019347469,0.0005004482,0.0006165694,0.0011411713,0.00020262809],"domain_scores_gemma":[0.9922518,0.0048819752,0.0005764014,0.0012971556,0.000823295,0.00016940644],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0047154073,0.0012954762,0.0005505838,0.003304435,0.0019361969,0.0065447814,0.0016401183,0.002352679,0.008925853],"category_scores_gemma":[0.012504252,0.000982642,0.0018228317,0.0031199192,0.007783898,0.021979246,0.0049127317,0.004860996,0.0026089721],"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":[0.000023735382,0.000012256358,0.00011181855,0.00021690363,0.000011611436,0.00022623409,0.0033514516,0.00077572383,0.0012466416,0.97698385,0.0023307933,0.0147090545],"study_design_scores_gemma":[0.000016260547,0.000025939913,0.00021370898,0.00017333803,0.00002016951,0.00035190495,0.0008820378,0.0038510142,0.0019797003,0.8887333,0.10371776,0.00003486634],"about_ca_topic_score_codex":0.0020458808,"about_ca_topic_score_gemma":0.0017654208,"teacher_disagreement_score":0.008925853,"about_ca_system_score_codex":0.0022038186,"about_ca_system_score_gemma":0.0016131089,"threshold_uncertainty_score":0.02985996},"labels":[],"label_agreement":null},{"id":"W2414577476","doi":"10.4242/balisagevol3.huitfeldt01","title":"Markup Meaning and Mereology","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Philosophy and Theoretical Science","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Mereology; Markup language; Meaning (existential); Computer science; Information retrieval; Epistemology; Philosophy; World Wide Web; XML","score_opus":0.013978655467464386,"score_gpt":0.27426816607449706,"score_spread":0.26028951060703265,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2414577476","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":[0.030244397,0.0155415265,0.6299571,0.017334444,0.0013541611,0.0001175361,0.0002270103,0.00063690596,0.3045869],"genre_scores_gemma":[0.8080956,0.006351582,0.15268804,0.002107688,0.001173954,0.00027207678,0.00017573618,0.00032824703,0.028807139],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99633324,0.0019088199,0.00024819537,0.0005597339,0.0007323146,0.00021760503],"domain_scores_gemma":[0.99555284,0.0026293383,0.000347,0.00091325294,0.00038632887,0.00017127844],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004433483,0.00055466784,0.0006094626,0.0025843976,0.00304779,0.008013851,0.0013145693,0.0025093774,0.0052727866],"category_scores_gemma":[0.00754557,0.000610268,0.0007801864,0.002180604,0.02897859,0.019338667,0.0044621164,0.003402168,0.0009190619],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000030216756,0.0000013996686,0.00002081942,0.000017156648,0.000001251624,0.000015844305,0.0006677617,0.00006719265,0.00008428899,0.9964289,0.00023489157,0.0024575284],"study_design_scores_gemma":[0.0000039891806,0.0000067828805,0.000059230184,0.000039357044,0.0000042022907,0.0000702637,0.00032287015,0.00033854978,0.000246047,0.97017854,0.028721852,0.000008259542],"about_ca_topic_score_codex":0.0011757419,"about_ca_topic_score_gemma":0.0008891498,"teacher_disagreement_score":0.008013851,"about_ca_system_score_codex":0.0029711258,"about_ca_system_score_gemma":0.001302395,"threshold_uncertainty_score":0.023446798},"labels":[],"label_agreement":null},{"id":"W2415938545","doi":"10.4242/balisagevol10.nordin01","title":"Markup and Canada's National Model Building Codes","year":2013,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Optics and Image Analysis","field":"Business, Management and Accounting","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Computer science; Markup language; Political science; World Wide Web; XML","score_opus":0.008138877558754322,"score_gpt":0.1904909705946607,"score_spread":0.18235209303590638,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2415938545","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":[0.017650194,0.004835141,0.05680211,0.036949016,0.0012997098,0.00031050717,0.005286836,0.002723737,0.87414277],"genre_scores_gemma":[0.34873217,0.011166704,0.15930687,0.006500936,0.00036083517,0.00055760564,0.007873112,0.0018495094,0.46365225],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.98562837,0.00083442143,0.00039918473,0.0005459545,0.011138109,0.0014539383],"domain_scores_gemma":[0.979677,0.0022715118,0.000619635,0.002065178,0.014292437,0.0010741596],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005539177,0.0006555223,0.0004042658,0.005346947,0.007562468,0.010194364,0.0030508328,0.001844732,0.015874827],"category_scores_gemma":[0.020135202,0.0005765686,0.0008102387,0.008325744,0.00472408,0.0036593527,0.0027484964,0.0023376746,0.0021881016],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000025742373,0.000017305916,0.00087841327,0.00011142781,0.0000071155746,0.000116970725,0.0012534178,0.003126649,0.00036811637,0.8316052,0.09462471,0.06786484],"study_design_scores_gemma":[0.0000105183835,0.000012680818,0.0022272028,0.00024642886,0.000011283674,0.000091298585,0.0007450078,0.0031312783,0.00086225,0.030380877,0.9622018,0.00007939151],"about_ca_topic_score_codex":0.98145884,"about_ca_topic_score_gemma":0.9800745,"teacher_disagreement_score":0.8984613,"about_ca_system_score_codex":0.10153872,"about_ca_system_score_gemma":0.18036993,"threshold_uncertainty_score":0.7367184},"labels":[],"label_agreement":null},{"id":"W2416130262","doi":"10.4242/balisagevol5.huitfeldt01","title":"Extension of the type/token distinction to document structure","year":2010,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","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":"Université de Montréal","funders":"","keywords":"Security token; Computer science; SGML; Document type definition; Extension (predicate logic); Type (biology); sort; Markup language; XML; Word (group theory); Information retrieval; Document Structure Description; Natural language processing; Linguistics; Programming language; World Wide Web","score_opus":0.016206641282411113,"score_gpt":0.21752586626501513,"score_spread":0.20131922498260402,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2416130262","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":[0.007483254,0.0005445273,0.9580768,0.0023283593,0.0006512715,0.00012893214,0.00043326564,0.0016912444,0.028662266],"genre_scores_gemma":[0.24856444,0.0013594441,0.7171831,0.0020344392,0.0010117167,0.00051967625,0.000823106,0.0017928617,0.026711198],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9914636,0.0030769133,0.0010709622,0.0015140093,0.0023042401,0.0005702839],"domain_scores_gemma":[0.9866518,0.00457765,0.00093774224,0.0054068533,0.001945402,0.00048052619],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005609986,0.0006639716,0.00075744634,0.0023353482,0.0016957483,0.007872285,0.0022813613,0.0019494764,0.0051646726],"category_scores_gemma":[0.012818843,0.0011143251,0.0019409041,0.0022116757,0.0071701678,0.019302052,0.004003878,0.004842424,0.0024965126],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000043091204,0.000015492888,0.000535182,0.00006744208,0.000010161569,0.0002499691,0.0015532551,0.0017431112,0.0016832775,0.9556666,0.0029076245,0.035524886],"study_design_scores_gemma":[0.00002971136,0.00006322684,0.0003670895,0.00013113981,0.00004022915,0.00089935656,0.00037611063,0.01172996,0.004604639,0.75087184,0.230816,0.00007080949],"about_ca_topic_score_codex":0.002436092,"about_ca_topic_score_gemma":0.0017169579,"teacher_disagreement_score":0.007872285,"about_ca_system_score_codex":0.0033241336,"about_ca_system_score_gemma":0.0023521427,"threshold_uncertainty_score":0.029668808},"labels":[],"label_agreement":null},{"id":"W2418003691","doi":"10.4242/balisagevol7.david01","title":"Challenges and Potential of Local Loading of XML Ebooks","year":2011,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Mathematics, Computing, and Information Processing","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Ontario Council of University Libraries; University of Toronto","funders":"","keywords":"Computer science; XML; World Wide Web; XML validation; Efficient XML Interchange; XSLT; XML Schema Editor; Presentation (obstetrics); XML framework; Upload; Streaming XML; Simple API for XML; Multimedia; XML Signature","score_opus":0.03435968104011919,"score_gpt":0.22018607522563133,"score_spread":0.18582639418551214,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2418003691","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":[0.32937217,0.004636565,0.39824197,0.03224203,0.0015481423,0.00093687855,0.0005634622,0.019614847,0.21284392],"genre_scores_gemma":[0.59908575,0.003940504,0.23273712,0.005324706,0.0015817407,0.0006579236,0.0011621306,0.016612718,0.13889742],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.98522687,0.007850321,0.00087429583,0.001108491,0.0040476643,0.00089235214],"domain_scores_gemma":[0.94863665,0.026046949,0.00204684,0.0122273,0.009205551,0.0018366601],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.017926661,0.00092123786,0.000887128,0.0024323782,0.0044254544,0.01150361,0.0043728803,0.0028444433,0.025283463],"category_scores_gemma":[0.057846773,0.0010169903,0.0007959821,0.0036365911,0.0040107598,0.019761387,0.009739494,0.0032056647,0.011865023],"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":[0.0005233614,0.00046541312,0.007956458,0.0017494461,0.000069073976,0.0044074156,0.12154198,0.0027200289,0.019940471,0.028114611,0.077502646,0.73500913],"study_design_scores_gemma":[0.00013384427,0.0007885174,0.00848367,0.0017530248,0.00014668533,0.0076996833,0.10860903,0.010809613,0.032660294,0.032603163,0.79586047,0.0004520782],"about_ca_topic_score_codex":0.0025257932,"about_ca_topic_score_gemma":0.0033681202,"teacher_disagreement_score":0.025283463,"about_ca_system_score_codex":0.0018536116,"about_ca_system_score_gemma":0.0021158,"threshold_uncertainty_score":0.09480637},"labels":[],"label_agreement":null},{"id":"W2471310903","doi":"10.4242/balisagevol9.zhao01","title":"Quality Control Practice for Scholars Portal, an XML-based E-journals Repository","year":2012,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Mathematics, Computing, and Information Processing","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":true,"ca_institutions":"Ontario Council of University Libraries","funders":"","keywords":"Computer science; Workflow; XML; World Wide Web; Information retrieval; Set (abstract data type); Quality (philosophy); Control (management); Database; Programming language","score_opus":0.03797933385401914,"score_gpt":0.33522051458186375,"score_spread":0.2972411807278446,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2471310903","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":[0.012307202,0.0049115024,0.84188217,0.026587807,0.0017180322,0.0072736004,0.003077882,0.016731774,0.08551006],"genre_scores_gemma":[0.11570498,0.003925528,0.8258437,0.004440153,0.0008392467,0.00492306,0.006480173,0.0037066508,0.03413652],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.81076616,0.0582817,0.024087643,0.013548673,0.09016191,0.0031539768],"domain_scores_gemma":[0.59889793,0.06046909,0.02424627,0.10559745,0.20495035,0.0058389013],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.19242798,0.0011974878,0.0013846403,0.011530922,0.0066078077,0.023989646,0.0077809775,0.003855105,0.00845255],"category_scores_gemma":[0.26557267,0.0013154761,0.0012131095,0.017765572,0.008802259,0.014253854,0.007694832,0.004249642,0.009293496],"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":[0.00033017385,0.00029454508,0.011153707,0.002650567,0.00020673321,0.00056273513,0.014481979,0.0030213005,0.00866315,0.27011013,0.12526105,0.56326383],"study_design_scores_gemma":[0.00013612003,0.00025754765,0.00873453,0.002564,0.000121681485,0.00066925655,0.0028561903,0.006100541,0.0151629085,0.040681772,0.9224787,0.00023675611],"about_ca_topic_score_codex":0.04018213,"about_ca_topic_score_gemma":0.027982367,"teacher_disagreement_score":0.9760104,"about_ca_system_score_codex":0.019538041,"about_ca_system_score_gemma":0.059990987,"threshold_uncertainty_score":0.99588007},"labels":[],"label_agreement":null},{"id":"W2474486350","doi":"10.4242/balisagevol1.marcoux01","title":"Graph characterization of overlap-only TexMECS and other overlapping markup formalisms","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Markup language; Computer science; Rotation formalisms in three dimensions; Graph; Programming language; Serialization; Serializability; Document Structure Description; Theoretical computer science; XML; Information retrieval; World Wide Web; Database transaction; Mathematics","score_opus":0.009800205548426531,"score_gpt":0.21770388659977685,"score_spread":0.20790368105135032,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2474486350","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":[0.07514876,0.00022887338,0.9083611,0.00046371119,0.000088554516,0.00013770231,0.0005424596,0.0018037001,0.0132250525],"genre_scores_gemma":[0.59010506,0.00037799272,0.3984532,0.00048974797,0.00014377022,0.00031421563,0.002378055,0.0010090534,0.0067288415],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9975439,0.0005132249,0.000241191,0.00068481785,0.0007403355,0.00027653578],"domain_scores_gemma":[0.9885832,0.004537875,0.0015342622,0.0029505182,0.0017620692,0.0006320619],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0019533508,0.00063410704,0.0005011148,0.0023372564,0.0013497479,0.0025587901,0.0014657457,0.0011405798,0.004042813],"category_scores_gemma":[0.008485476,0.0004982425,0.001458793,0.0018089681,0.0034922762,0.005917287,0.0026162534,0.0019513236,0.00092308386],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00017468563,0.00006085807,0.0014498951,0.00026760346,0.000032773816,0.0008479308,0.0010792497,0.015543552,0.013506527,0.9377611,0.0018236794,0.027452182],"study_design_scores_gemma":[0.000054976917,0.000117249976,0.0013652217,0.00014251212,0.000113656555,0.0014519264,0.000781343,0.089826696,0.032568693,0.81892264,0.054568194,0.000086817585],"about_ca_topic_score_codex":0.003215521,"about_ca_topic_score_gemma":0.0026733277,"teacher_disagreement_score":0.004042813,"about_ca_system_score_codex":0.0011477083,"about_ca_system_score_gemma":0.0014145466,"threshold_uncertainty_score":0.013524532},"labels":[],"label_agreement":null},{"id":"W2499820415","doi":"10.4242/balisagevol6.ruus01","title":"A brief history of markup of social science data: from punched cards to \"the life cycle\" approach","year":2010,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Research Data Management Practices","field":"Computer Science","cited_by":0,"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 Toronto","funders":"","keywords":"Metadata; Computer science; Markup language; Raw data; Documentation; Software; Metadata repository; Process (computing); Generator (circuit theory); World Wide Web; Data science; Information retrieval; XML","score_opus":0.06953390648258413,"score_gpt":0.3141411563154728,"score_spread":0.24460724983288867,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2499820415","genre_codex":"review","genre_gemma":"review","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":"review","domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004728328,0.33225524,0.30946362,0.090853736,0.008963626,0.0005326905,0.0016925843,0.0013152934,0.2501948],"genre_scores_gemma":[0.080026776,0.47097585,0.309338,0.02578567,0.010943004,0.0016757803,0.0024110957,0.0016053236,0.09723853],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9928404,0.0041045514,0.00074096,0.0006418116,0.0014572799,0.00021496067],"domain_scores_gemma":[0.97828126,0.017485244,0.00065898243,0.0016009442,0.0015919637,0.0003815342],"candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.009583402,0.0008195337,0.0006964736,0.010635022,0.0033199403,0.010194489,0.0021475744,0.00343264,0.010220413],"category_scores_gemma":[0.020284785,0.0010543213,0.000726955,0.016086556,0.01354335,0.01771965,0.0033949616,0.0068382476,0.005490147],"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":[0.000036125806,0.000046521887,0.0006585747,0.0011069197,0.000014822765,0.00023419921,0.004192096,0.00049294165,0.00043101006,0.7081663,0.047441714,0.23717888],"study_design_scores_gemma":[0.0000027960261,0.000026910964,0.0004201822,0.00093615183,0.000004308995,0.00025065738,0.00084716553,0.00039407227,0.00051481207,0.100925736,0.8956467,0.00003052129],"about_ca_topic_score_codex":0.0049358467,"about_ca_topic_score_gemma":0.0046366937,"teacher_disagreement_score":0.9904166,"about_ca_system_score_codex":0.0067885495,"about_ca_system_score_gemma":0.0033962314,"threshold_uncertainty_score":0.050682485},"labels":[],"label_agreement":null},{"id":"W2500534248","doi":"10.4242/balisagevol15.marcoux01","title":"Applying intertextual semantics to Cyberjustice","year":2015,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Artificial Intelligence in Law","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science; nobody; Semantics (computer science); Process (computing); Forcing (mathematics); Project management; Engineering; Systems engineering; Programming language","score_opus":0.09485323916849173,"score_gpt":0.3652355035328645,"score_spread":0.27038226436437274,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2500534248","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":[0.04408714,0.0010146882,0.88605344,0.009786549,0.00028139923,0.00014650417,0.00013694515,0.0007223509,0.05777093],"genre_scores_gemma":[0.70344406,0.0013468924,0.2807891,0.0011116399,0.00022823102,0.0002450215,0.0003334333,0.00057428505,0.011927294],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.990576,0.006247329,0.00037957449,0.0006969545,0.0016329149,0.00046723065],"domain_scores_gemma":[0.9856812,0.008583957,0.0007632273,0.0034619803,0.0010930954,0.00041663705],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.008544994,0.00081718765,0.00062735664,0.0024389268,0.0024896385,0.0069279973,0.0020959617,0.002476333,0.005764566],"category_scores_gemma":[0.019050881,0.0005704875,0.0012179977,0.0021484166,0.012061414,0.015706385,0.007427107,0.004395983,0.00059575075],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000011004006,0.000037740425,0.00062213803,0.00006945109,0.000013123581,0.00015353304,0.0039906944,0.007412565,0.0003435422,0.9716789,0.0009552477,0.014712132],"study_design_scores_gemma":[0.00001367025,0.00004264653,0.00041779436,0.00013112363,0.000020061587,0.00021923,0.0033590687,0.034225337,0.0029661846,0.8665058,0.09206754,0.00003154671],"about_ca_topic_score_codex":0.012621751,"about_ca_topic_score_gemma":0.012755098,"teacher_disagreement_score":0.012621751,"about_ca_system_score_codex":0.0047053723,"about_ca_system_score_gemma":0.0040902486,"threshold_uncertainty_score":0.04519081},"labels":[],"label_agreement":null},{"id":"W2516342496","doi":"10.4242/balisagevol10.marcoux01","title":"Modeling overlapping structures","year":2013,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Computer science","score_opus":0.03674893105993818,"score_gpt":0.21381884489443073,"score_spread":0.17706991383449255,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2516342496","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":[0.029910108,0.00041981624,0.9546604,0.0004606325,0.00005142987,0.00019487069,0.0006580914,0.0012750782,0.012369525],"genre_scores_gemma":[0.32826242,0.0006928083,0.65614396,0.00020474315,0.0001071179,0.00037438702,0.002758224,0.000647495,0.010808961],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9972045,0.00070642174,0.00017289996,0.0008293418,0.0007411798,0.00034574533],"domain_scores_gemma":[0.9868728,0.0070824637,0.001128886,0.0031333165,0.0011650125,0.0006175217],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021065741,0.0010167083,0.0008036371,0.0023123415,0.0020863593,0.0035539155,0.0035254667,0.0020030502,0.007894964],"category_scores_gemma":[0.01575293,0.0011753433,0.001611088,0.0032417187,0.0025150413,0.010786139,0.0051902914,0.0024483406,0.0016166883],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013976874,0.00009059156,0.002296819,0.00025000327,0.000039255505,0.00072214496,0.0014100152,0.11141536,0.0032157619,0.81956553,0.005687509,0.05516733],"study_design_scores_gemma":[0.00002662137,0.0000360678,0.0003233354,0.00006261505,0.000032633427,0.00037600467,0.0004259781,0.27894002,0.0031799595,0.6811778,0.03539429,0.000024694087],"about_ca_topic_score_codex":0.005873424,"about_ca_topic_score_gemma":0.006981112,"teacher_disagreement_score":0.007894964,"about_ca_system_score_codex":0.0022506113,"about_ca_system_score_gemma":0.0017671519,"threshold_uncertainty_score":0.026411295},"labels":[],"label_agreement":null},{"id":"W2522795258","doi":"10.4242/balisagevol10.cameron01","title":"icXML: Accelerating a Commercial XML Parser Using SIMD and Multicore Technologies","year":2013,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Advanced Database Systems and Queries","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":"Simon Fraser University","funders":"","keywords":"Computer science; SIMD; XML; Parsing; Pipeline (software); Simple API for XML; Programming language; XML framework; Multi-core processor; Parallel computing; Software engineering; XML Signature; Operating system; Efficient XML Interchange","score_opus":0.032674047108623594,"score_gpt":0.25233480553892,"score_spread":0.21966075843029642,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2522795258","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":[0.0413374,0.00087265985,0.744373,0.0010242754,0.00042912713,0.0006576342,0.0021573084,0.18394819,0.025200477],"genre_scores_gemma":[0.09749367,0.00049395947,0.86611414,0.0007246871,0.00008350095,0.00036713938,0.006767032,0.010067682,0.017888194],"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99847203,0.00021312862,0.00011163952,0.0002806925,0.00077434233,0.00014806415],"domain_scores_gemma":[0.9977254,0.0005461783,0.000156966,0.0006251706,0.0008168743,0.00012932919],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027505897,0.00090536976,0.0004450667,0.0011034639,0.00059382844,0.0019887895,0.0028755192,0.0009820533,0.010206964],"category_scores_gemma":[0.0044083633,0.00072139164,0.0006397155,0.0015526805,0.0006067259,0.0034765918,0.001317626,0.001602887,0.0039887438],"study_design_candidate":"bench_or_experimental","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":[0.0014937938,0.00047585825,0.0054776357,0.0010589847,0.00019955635,0.0010407721,0.001267365,0.016688215,0.1755327,0.05292619,0.14770898,0.59612995],"study_design_scores_gemma":[0.0005980068,0.0006288977,0.0035828096,0.00017400691,0.00009538662,0.0010771403,0.0002866265,0.32059982,0.3600838,0.012566847,0.30008146,0.00022523582],"about_ca_topic_score_codex":0.004285934,"about_ca_topic_score_gemma":0.0048208027,"teacher_disagreement_score":0.010206964,"about_ca_system_score_codex":0.0014111928,"about_ca_system_score_gemma":0.0022391246,"threshold_uncertainty_score":0.034145713},"labels":[],"label_agreement":null},{"id":"W2756031164","doi":"10.4242/balisagevol19.lee01","title":"The Secret Life of Schema in Web Protocols, API's and Software Type Systems","year":2017,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Service-Oriented Architecture and Web Services","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Nexen (Canada)","funders":"","keywords":"Computer science; JSON; Document Structure Description; Software engineering; Programming language; XML; Software; Markup language; Schema (genetic algorithms); World Wide Web; Information retrieval","score_opus":0.016728122882226866,"score_gpt":0.2596243976694916,"score_spread":0.24289627478726472,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2756031164","genre_codex":"methods","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.01805213,0.010507635,0.8516585,0.032002315,0.00267707,0.00016991187,0.0002760152,0.0011248267,0.08353159],"genre_scores_gemma":[0.41303015,0.01576064,0.5226289,0.009081624,0.0028183844,0.0009008785,0.0005018603,0.0025706752,0.03270696],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9851517,0.007880216,0.00089349877,0.001263452,0.004298648,0.0005125043],"domain_scores_gemma":[0.9596669,0.022987906,0.0014500904,0.012846182,0.0023567872,0.000692198],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.021896327,0.0006021994,0.0009474303,0.0023825807,0.004401162,0.017903354,0.0020317787,0.0043293065,0.005918375],"category_scores_gemma":[0.04208486,0.0012359733,0.00095311075,0.0040081074,0.033834822,0.049417514,0.0057958844,0.011279807,0.0019746658],"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":[0.000013807936,0.000010195548,0.00013468685,0.00004345687,0.000005162877,0.00003625248,0.0014952169,0.00025062208,0.00027535346,0.9864509,0.0012126702,0.010071584],"study_design_scores_gemma":[0.00001536048,0.000048948128,0.00016013224,0.00021324705,0.000014065019,0.00032426766,0.001145772,0.0021732983,0.0029594833,0.8450095,0.14789304,0.000042876767],"about_ca_topic_score_codex":0.001579339,"about_ca_topic_score_gemma":0.0009993427,"teacher_disagreement_score":0.021896327,"about_ca_system_score_codex":0.0030970294,"about_ca_system_score_gemma":0.0030433205,"threshold_uncertainty_score":0.11580026},"labels":[],"label_agreement":null},{"id":"W2941403856","doi":"10.4242/balisagevol22.morrissey01","title":"“Be in the Room Where It Happens”","year":2018,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Diverse academic research themes","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Nuclear Physics; National Institute of Information and Communications Technology; Japan Aerospace Exploration Agency; Canadian Space Agency; Center for Long-Term Cybersecurity, University of California Berkeley; Geo-Informatics and Space Technology Development Agency; Iran Telecommunication Research Center; European Space Agency; Agenzia Spaziale Italiana; European Organization for the Exploitation of Meteorological Satellites; Korea Aerospace Research Institute; National Commission for Science and Technology; Türkiye Bilimsel ve Teknolojik Araştırma Kurumu; National Oceanic and Atmospheric Administration; National Space Organization; China National Space Administration; Belgian Federal Science Policy Office; Chinese Academy of Sciences; U.S. Geological Survey; Indian Space Research Organisation; Centre National d’Etudes Spatiales; Commonwealth Scientific and Industrial Research Organisation; National Aeronautics and Space Administration","keywords":"USable; Verifiable secret sharing; Computer science; Key (lock); Task (project management); Face (sociological concept); Resource (disambiguation); Digital content; Internet privacy; Data science; World Wide Web; Multimedia; Computer security; Engineering; Sociology; Set (abstract data type)","score_opus":0.16508507791108423,"score_gpt":0.41564468531797755,"score_spread":0.2505596074068933,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2941403856","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.016483855,0.0022057695,0.056944773,0.16504532,0.0049127927,0.0003391262,0.0002392691,0.0010242903,0.75280476],"genre_scores_gemma":[0.4490468,0.004799813,0.045759328,0.046710726,0.0013224081,0.00068392407,0.000478294,0.0009012267,0.45029745],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9913496,0.0050505423,0.00020443113,0.00088563346,0.0016599722,0.0008498895],"domain_scores_gemma":[0.99113715,0.0019664061,0.0008240109,0.0016125543,0.0016762195,0.002783722],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0073003992,0.00080304797,0.00044179754,0.0011273854,0.011644704,0.014300143,0.0016579697,0.004572234,0.037142556],"category_scores_gemma":[0.0153125,0.00041468718,0.00056817284,0.0012884781,0.01851664,0.017317777,0.008981048,0.005860711,0.020176202],"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":[0.00011543516,0.00014528386,0.0025815165,0.0004304707,0.00003098046,0.00074331975,0.05627034,0.00030159234,0.002123116,0.5254949,0.319651,0.09211202],"study_design_scores_gemma":[0.000006403991,0.00003258036,0.0004443331,0.00027220132,0.000008659743,0.00028970925,0.031627867,0.0001296691,0.0005382487,0.03277292,0.9338398,0.00003768921],"about_ca_topic_score_codex":0.004142485,"about_ca_topic_score_gemma":0.009109917,"teacher_disagreement_score":0.037142556,"about_ca_system_score_codex":0.0030722222,"about_ca_system_score_gemma":0.006161267,"threshold_uncertainty_score":0.12425417},"labels":[],"label_agreement":null},{"id":"W2979338130","doi":"10.4242/balisagevol23.boyer01","title":"Multitasking algorithms in XForms","year":2019,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"IBM (Canada)","funders":"","keywords":"Human multitasking; Computer science; Algorithm","score_opus":0.025212435694555068,"score_gpt":0.21830197535124637,"score_spread":0.19308953965669132,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2979338130","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":[0.013779919,0.0003039856,0.97461176,0.00017402422,0.000056401237,0.000087216606,0.00008126854,0.004396565,0.006508934],"genre_scores_gemma":[0.1416397,0.00021506002,0.8473297,0.00012171675,0.00006253456,0.00023728072,0.00034008542,0.0007573579,0.009296563],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99813676,0.00039547944,0.00027880678,0.0003640009,0.0005936654,0.00023137825],"domain_scores_gemma":[0.99763477,0.001170549,0.0001821822,0.00058366003,0.00033652672,0.00009221606],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0029816548,0.0006201208,0.00068789016,0.00092657947,0.0008753155,0.002087979,0.0019132822,0.0010150679,0.007348956],"category_scores_gemma":[0.006367808,0.00042680514,0.00095563906,0.0009092908,0.0010485947,0.0026247045,0.0018978205,0.0009956596,0.0016781371],"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":[0.0005209455,0.00019341985,0.00207997,0.0003505987,0.000060730934,0.00027297132,0.00072886614,0.18485922,0.014761612,0.30778793,0.009967607,0.47841603],"study_design_scores_gemma":[0.00012853986,0.00010539022,0.00052405946,0.00008664171,0.000020411486,0.00021458534,0.00007848048,0.8052003,0.0181876,0.14015964,0.035251692,0.000042690306],"about_ca_topic_score_codex":0.0036507277,"about_ca_topic_score_gemma":0.0032168848,"teacher_disagreement_score":0.007348956,"about_ca_system_score_codex":0.0014774934,"about_ca_system_score_gemma":0.0013005338,"threshold_uncertainty_score":0.02458471},"labels":[],"label_agreement":null},{"id":"W4205261768","doi":"10.4242/balisagevol13.sperberg-mcqueen01","title":"Document lattices","year":2014,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Markup language; XHTML; RuleML; Computer science; Document type definition; Information retrieval; Consistency (knowledge bases); Document Structure Description; XML; Programming language; Natural language processing; World Wide Web; Artificial intelligence","score_opus":0.025126128682454143,"score_gpt":0.22030216916037387,"score_spread":0.19517604047791973,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4205261768","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":[0.0082973745,0.0019220748,0.86236984,0.002248431,0.0007852328,0.00055512897,0.009642805,0.0056682588,0.10851086],"genre_scores_gemma":[0.1488097,0.0024137679,0.753168,0.0008895657,0.0006751714,0.0009868452,0.017767254,0.0018163649,0.07347328],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9902844,0.0024800233,0.0014813514,0.002028455,0.0032468203,0.00047895304],"domain_scores_gemma":[0.9824547,0.006995213,0.0008125764,0.0046299035,0.0039160224,0.0011915475],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0055440827,0.0009298698,0.0011123716,0.0042495113,0.003515126,0.011462895,0.0030373328,0.0014186983,0.038777284],"category_scores_gemma":[0.02024767,0.0010392829,0.0023863749,0.0048311837,0.00372422,0.013222847,0.006109642,0.0032262942,0.018192872],"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":[0.00011090213,0.000042455293,0.0002385924,0.00023150476,0.00002259321,0.0001679053,0.00070963654,0.0017465843,0.0010278533,0.9211155,0.014746584,0.059839927],"study_design_scores_gemma":[0.00005674799,0.000035911748,0.00012227664,0.00011302832,0.000019233266,0.00047832448,0.00047193826,0.0066550565,0.0025211265,0.7100438,0.27943426,0.000048271933],"about_ca_topic_score_codex":0.003080277,"about_ca_topic_score_gemma":0.0024754675,"teacher_disagreement_score":0.038777284,"about_ca_system_score_codex":0.0025597129,"about_ca_system_score_gemma":0.0029168136,"threshold_uncertainty_score":0.12972295},"labels":[],"label_agreement":null},{"id":"W4206983357","doi":"10.4242/balisagevol7.usdin01","title":"Serendipity","year":2011,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Serendipity; Luck; Face (sociological concept); Key (lock); Computer science; Psychology; Sociology; Epistemology; Computer security; Social science","score_opus":0.7067267809977053,"score_gpt":0.5319341951220168,"score_spread":0.17479258587568847,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4206983357","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":[0.0021509477,0.0036240295,0.015911734,0.019161338,0.034225535,0.0003012734,0.0025655539,0.006729361,0.91533023],"genre_scores_gemma":[0.032790117,0.0031732018,0.0063934163,0.0066683153,0.015336183,0.00049549417,0.0028485807,0.005846734,0.9264479],"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9906194,0.0025223934,0.00055814395,0.0016518055,0.0033939227,0.0012543134],"domain_scores_gemma":[0.96956885,0.0047531915,0.0018987786,0.0076103304,0.008678655,0.0074901422],"candidate_categories":["bibliometrics","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00852282,0.0015323645,0.0012746316,0.005306947,0.004185578,0.019279124,0.0027939943,0.0028542818,0.4072721],"category_scores_gemma":[0.040621664,0.0007804954,0.0011917969,0.003196487,0.0029072438,0.012361515,0.014918629,0.004399833,0.30046523],"study_design_candidate":"theoretical_or_conceptual","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":[0.000112625334,0.00002775957,0.00030818742,0.00038853058,0.000015796651,0.00023187058,0.00078405096,0.00009276243,0.0007646117,0.074919164,0.8279848,0.09436986],"study_design_scores_gemma":[0.0000066745606,0.000017140734,0.000117498486,0.00009346102,0.000003887434,0.00022545009,0.000106664724,0.000045904984,0.0002382579,0.0047472687,0.9943878,0.000009973217],"about_ca_topic_score_codex":0.0005634092,"about_ca_topic_score_gemma":0.00076054013,"teacher_disagreement_score":0.99469304,"about_ca_system_score_codex":0.002397399,"about_ca_system_score_gemma":0.0038858813,"threshold_uncertainty_score":0.8454544},"labels":[],"label_agreement":null},{"id":"W4230652012","doi":"10.4242/balisagevol26.jenstad01","title":"Converting an SGML/XML Hybrid to TEI-XML: The Case of the Internet Shakespeare Editions","year":2021,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":1,"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":"SGML; Markup language; XML; Computer science; The Internet; World Wide Web; RuleML; Encoding (memory); Document type definition; Software; Code (set theory); Programming language; Document Structure Description; Artificial intelligence","score_opus":0.03736616915092719,"score_gpt":0.23516744469880627,"score_spread":0.19780127554787907,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4230652012","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.46979764,0.0014964325,0.2123682,0.02525488,0.001658242,0.0003870526,0.00073801796,0.0045198672,0.28377977],"genre_scores_gemma":[0.6840191,0.0009327843,0.1677799,0.003338192,0.00023373477,0.00015481445,0.0008242668,0.0049284254,0.13778883],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9963581,0.0013962158,0.00019016088,0.00035997547,0.0013982479,0.00029728594],"domain_scores_gemma":[0.9899037,0.004298465,0.00039535842,0.003255675,0.0016921642,0.00045467343],"candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.008568699,0.00030753485,0.00021941694,0.0015861049,0.004963496,0.007471955,0.0015025816,0.0028204462,0.0048948126],"category_scores_gemma":[0.023489932,0.000528213,0.00036597482,0.001920344,0.00630954,0.008221885,0.0048241396,0.003607375,0.00207588],"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":[0.00019412396,0.0002084677,0.0065444745,0.00028780792,0.00002755709,0.015241089,0.15658422,0.0050757634,0.009519082,0.54648715,0.051009938,0.20882031],"study_design_scores_gemma":[0.000037448353,0.00009701921,0.0030142097,0.00053601817,0.000028033559,0.0037205473,0.03230056,0.00862414,0.014145779,0.060684856,0.87667125,0.00014005449],"about_ca_topic_score_codex":0.013776188,"about_ca_topic_score_gemma":0.022020832,"teacher_disagreement_score":0.992528,"about_ca_system_score_codex":0.0027222019,"about_ca_system_score_gemma":0.0024032167,"threshold_uncertainty_score":0.04531616},"labels":[],"label_agreement":null},{"id":"W4255233508","doi":"10.4242/balisagevol1.levy01","title":"Beyond the Semantic Web: the Semantic Space","year":2009,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Semantic Web and Ontologies","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":"University of Ottawa","funders":"","keywords":"Computer science; Parsing; Semantics (computer science); Semantic Web; Semantic Web Rule Language; Space (punctuation); Social Semantic Web; Variety (cybernetics); Natural language; Artificial intelligence; World Wide Web; Natural language processing; Semantic analytics; Programming language","score_opus":0.009949442858470403,"score_gpt":0.2247738060855085,"score_spread":0.2148243632270381,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4255233508","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":[0.011241459,0.077315755,0.48626667,0.14461991,0.0052759256,0.00024298948,0.0016150916,0.0026732117,0.27074897],"genre_scores_gemma":[0.4326375,0.09163036,0.37365994,0.029180728,0.008436347,0.0007314469,0.0038352276,0.0018986416,0.05798982],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9886431,0.005617003,0.00065312255,0.0012252345,0.0032177877,0.0006437129],"domain_scores_gemma":[0.98742247,0.006740516,0.00052166433,0.00294762,0.0013348291,0.0010329239],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01009487,0.001158196,0.001787826,0.0059225988,0.0046946476,0.02644896,0.003664196,0.0072986693,0.0085127195],"category_scores_gemma":[0.014970833,0.00083125767,0.0015030168,0.009781922,0.020236181,0.08037509,0.009022722,0.008563944,0.0032110494],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00001898257,0.00001273398,0.00012821871,0.00013401131,0.00001637107,0.00009876644,0.00075472787,0.0005274978,0.00017602388,0.97272235,0.0081832325,0.01722708],"study_design_scores_gemma":[0.000005494919,0.0000068067143,0.00006335042,0.00019193672,0.000007538348,0.00010950533,0.00065018417,0.0014123282,0.00016068864,0.8255877,0.17178865,0.000015884496],"about_ca_topic_score_codex":0.0053295353,"about_ca_topic_score_gemma":0.0024852126,"teacher_disagreement_score":0.02644896,"about_ca_system_score_codex":0.0054961275,"about_ca_system_score_gemma":0.006833602,"threshold_uncertainty_score":0.053387403},"labels":[],"label_agreement":null},{"id":"W4288685011","doi":"10.4242/balisagevol27.takeda01","title":"Serverless Searching with XSLT and JavaScript","year":2022,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Web Data Mining and Analysis","field":"Computer Science","cited_by":1,"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; Simon Fraser University","funders":"","keywords":"XSLT; JavaScript; Computer science; World Wide Web; Bytecode; Ajax; Presentation (obstetrics); XML; Information retrieval; Programming language; Web application; Virtual machine","score_opus":0.008208968830791925,"score_gpt":0.19156553632581513,"score_spread":0.18335656749502321,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4288685011","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":[0.02532064,0.0010548491,0.7739262,0.0005981016,0.00025547808,0.00026827265,0.0044709635,0.15939508,0.03471039],"genre_scores_gemma":[0.19320512,0.0013722718,0.7188498,0.0008219897,0.00026510956,0.0002898755,0.016152253,0.024954144,0.04408952],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.99648273,0.0005050178,0.00049074326,0.0006890154,0.0016025505,0.0002299304],"domain_scores_gemma":[0.9930307,0.0021822508,0.00044329235,0.0029305385,0.0011744802,0.00023870435],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0026357258,0.00093843654,0.0012510213,0.0038421294,0.0012036477,0.0053816293,0.0017966016,0.0010543303,0.010656653],"category_scores_gemma":[0.010877838,0.0009100044,0.0012104794,0.0049096486,0.0011996039,0.007909952,0.0026418704,0.0012712048,0.013491408],"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":[0.00093376567,0.00035271028,0.006141546,0.000924097,0.00018590516,0.0008236925,0.0019967065,0.0048487657,0.075623736,0.051057704,0.102525875,0.7545855],"study_design_scores_gemma":[0.00020114142,0.00038866984,0.006773272,0.00054677937,0.0002024603,0.0020523078,0.00073193625,0.14208351,0.2519754,0.07975338,0.5148934,0.00039778068],"about_ca_topic_score_codex":0.0032691967,"about_ca_topic_score_gemma":0.0040193903,"teacher_disagreement_score":0.010656653,"about_ca_system_score_codex":0.00090663315,"about_ca_system_score_gemma":0.002129517,"threshold_uncertainty_score":0.035650074},"labels":[],"label_agreement":null},{"id":"W4288685812","doi":"10.4242/balisagevol27.holstege01","title":"Metadata for Creators","year":2022,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Library Science and Information Systems","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Nuclear Physics; National Institute of Information and Communications Technology; Japan Aerospace Exploration Agency; Canadian Space Agency; Center for Long-Term Cybersecurity, University of California Berkeley; Geo-Informatics and Space Technology Development Agency; Iran Telecommunication Research Center; European Space Agency; Agenzia Spaziale Italiana; European Organization for the Exploitation of Meteorological Satellites; Korea Aerospace Research Institute; National Commission for Science and Technology; Türkiye Bilimsel ve Teknolojik Araştırma Kurumu; National Oceanic and Atmospheric Administration; National Space Organization; China National Space Administration; Belgian Federal Science Policy Office; Chinese Academy of Sciences; U.S. Geological Survey; Indian Space Research Organisation; Centre National d’Etudes Spatiales; Commonwealth Scientific and Industrial Research Organisation; National Aeronautics and Space Administration","keywords":"Metadata; Computer science; Meta Data Services; World Wide Web; Geospatial metadata; Process (computing); Key (lock); Metadata repository; Data element; Information retrieval; Computer security","score_opus":0.018163097642529862,"score_gpt":0.21971845434499412,"score_spread":0.20155535670246427,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4288685812","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":[0.0083025275,0.006603123,0.5696186,0.029225161,0.0052487073,0.0012088922,0.003956476,0.007711923,0.36812466],"genre_scores_gemma":[0.14280517,0.011484958,0.44644657,0.0075038197,0.0033756546,0.001600833,0.008674836,0.007844023,0.3702641],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.98666734,0.005031167,0.0016390906,0.001643293,0.0042989864,0.00072022557],"domain_scores_gemma":[0.9371588,0.013796139,0.0031886678,0.031474497,0.011931515,0.0024503989],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.015276614,0.00091947656,0.00084780325,0.0088078035,0.0062237806,0.017648306,0.0034097089,0.0042315535,0.04528815],"category_scores_gemma":[0.06596929,0.0010813815,0.0010643831,0.007802503,0.005745183,0.040574946,0.013292292,0.0052039195,0.027653689],"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":[0.00008500395,0.000047703994,0.0013349666,0.00051882566,0.000017914828,0.0003019638,0.009432956,0.0004917587,0.0029778609,0.6881008,0.065114014,0.23157626],"study_design_scores_gemma":[0.000008562749,0.000019492973,0.00022422275,0.00045494636,0.0000109892935,0.00027699117,0.001249468,0.00032943435,0.0017409567,0.054349788,0.9412996,0.000035522626],"about_ca_topic_score_codex":0.0042253756,"about_ca_topic_score_gemma":0.0035816417,"teacher_disagreement_score":0.04528815,"about_ca_system_score_codex":0.0043034153,"about_ca_system_score_gemma":0.008223115,"threshold_uncertainty_score":0.15150392},"labels":[],"label_agreement":null},{"id":"W4288686143","doi":"10.4242/balisagevol27.david01","title":"BITS for Government Information?","year":2022,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Mathematics, Computing, and Information Processing","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Ontario Council of University Libraries; University of Toronto","funders":"","keywords":"Metadata; Suite; Computer science; World Wide Web; Schema (genetic algorithms); Government (linguistics); Rewriting; Vocabulary; Information retrieval; Political science","score_opus":0.008548217516537648,"score_gpt":0.20715457783906724,"score_spread":0.1986063603225296,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4288686143","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":[0.0020405739,0.0076543167,0.005983101,0.16864645,0.014652316,0.00007594221,0.002851397,0.001960991,0.7961349],"genre_scores_gemma":[0.05122243,0.0127236685,0.010971538,0.08514365,0.004507915,0.00021791922,0.004048864,0.0031170228,0.8280469],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99632984,0.0010389183,0.00025877453,0.0003045697,0.0015623239,0.0005055231],"domain_scores_gemma":[0.9938,0.0010960416,0.0003897256,0.0020173427,0.0017445048,0.000952401],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0027735573,0.0004733358,0.00044485382,0.002519891,0.003484219,0.0153114,0.0011650582,0.0048830807,0.19835666],"category_scores_gemma":[0.01672762,0.00044153965,0.00039987487,0.005498972,0.0048006815,0.024405383,0.0069585745,0.0040841694,0.11698899],"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":[0.00002395085,0.0000091647435,0.00024274521,0.000099315934,0.0000034419747,0.00007525822,0.00083950325,0.000028193883,0.00024437348,0.27809322,0.66891456,0.051426318],"study_design_scores_gemma":[0.0000013511582,0.0000018899,0.000064987566,0.000049690552,9.315423e-7,0.000028548928,0.0002745745,0.000008077407,0.00006467423,0.0052745612,0.99422693,0.000003753987],"about_ca_topic_score_codex":0.012932824,"about_ca_topic_score_gemma":0.01653783,"teacher_disagreement_score":0.19835666,"about_ca_system_score_codex":0.0047905203,"about_ca_system_score_gemma":0.0044427407,"threshold_uncertainty_score":0.6635691},"labels":[],"label_agreement":null},{"id":"W4385386714","doi":"10.4242/balisagevol28.hymers01","title":"Retractions and Corrections at Scholars Portal Journals","year":2023,"lang":"en","type":"article","venue":"Balisage series on markup technologies","topic":"Library Science and Information Systems","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Ontario Council of University Libraries","funders":"","keywords":"Metadata; Computer science; Workflow; Identifier; World Wide Web; XML; Process (computing); Digital library; Service (business); Object (grammar); Information retrieval; Library science; Database; Business","score_opus":0.025519186116147733,"score_gpt":0.2562568147893987,"score_spread":0.23073762867325098,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4385386714","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"evaluation","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"evaluation","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.05742228,0.011465177,0.08566204,0.1469542,0.17899732,0.0060721356,0.012466904,0.042452097,0.45850787],"genre_scores_gemma":[0.15892091,0.013007359,0.17050922,0.020505529,0.032386143,0.0019066238,0.0106553435,0.028236838,0.56387204],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.89193904,0.022577496,0.020538202,0.00525962,0.05615456,0.0035310576],"domain_scores_gemma":[0.47482985,0.13097622,0.054127377,0.10909099,0.21169701,0.019278534],"candidate_categories":["metaresearch","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.07890017,0.0014368315,0.0018236736,0.0274396,0.01155584,0.021018686,0.0044684056,0.0043889554,0.064721994],"category_scores_gemma":[0.45327607,0.0018743237,0.0024624437,0.019827956,0.0052583646,0.01201517,0.012848752,0.0071132337,0.0660854],"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":[0.00046149068,0.00016981219,0.0063364753,0.0018052771,0.000101056015,0.0028292683,0.009805723,0.0003432115,0.006038383,0.023678388,0.6437863,0.30464458],"study_design_scores_gemma":[0.00003649776,0.000100044446,0.0026981093,0.0006870707,0.00003919228,0.00059503387,0.0024859444,0.00034750978,0.004319283,0.006348402,0.98218495,0.00015796191],"about_ca_topic_score_codex":0.0091172075,"about_ca_topic_score_gemma":0.016838834,"teacher_disagreement_score":0.9956111,"about_ca_system_score_codex":0.0077890917,"about_ca_system_score_gemma":0.033500925,"threshold_uncertainty_score":0.417269},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"}]}