{"id":"W4391464262","doi":"10.1162/qss_a_00285","title":"Large-scale text analysis using generative language models: A case study in discovering public value expressions in AI patents","year":2024,"lang":"en","type":"article","venue":"Quantitative Science Studies","topic":"Computational and Text Analysis Methods","field":"Social Sciences","cited_by":30,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; Biotechnology and Biological Sciences Research Council; Directorate for Biological Sciences; Snap","keywords":"Generative grammar; Natural language processing; Scale (ratio); Value (mathematics); Computer science; Generative model; Artificial intelligence; Linguistics; Machine learning; Geography; Cartography; Philosophy","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004589079,0.0001581162,0.000400745,0.001667974,0.001162907,0.0003684238,0.0002608649,0.00002556232,0.00001683716],"category_scores_gemma":[0.0008308379,0.0001246912,0.000123847,0.01105967,0.000752338,0.001719394,0.0002850913,0.0001847978,0.000003557703],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000386888,"about_ca_system_score_gemma":0.0003830986,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.009356229,"about_ca_topic_score_gemma":0.112171,"domain_scores_codex":[0.9967972,0.0009426446,0.0003689601,0.0006520539,0.0007587001,0.0004804506],"domain_scores_gemma":[0.9987276,0.0007081497,0.00007110113,0.0001383552,0.0002645648,0.00009020454],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.000003841752,0.0002570648,0.03065536,0.000008849272,0.0002860525,0.0006471293,0.8510596,0.0891554,0.0003903407,0.02659447,0.000003172498,0.0009387207],"study_design_scores_gemma":[0.00009261802,0.00002336919,0.002632314,0.00003627691,0.00008065416,0.000001369845,0.57946,0.4135927,0.00001121583,0.003951295,0.000005198558,0.0001130049],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9467995,0.002240204,0.04970578,0.0003653085,0.0001667548,0.0002688332,0.00001482843,0.00003366812,0.0004051544],"genre_scores_gemma":[0.9898112,0.00006019706,0.00991893,0.00003651321,0.00003065304,0.00004214632,0.000001285286,0.000006243813,0.00009277736],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3244373,"threshold_uncertainty_score":0.9972405,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2553169436416557,"score_gpt":0.5238114656364624,"score_spread":0.2684945219948067,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}