{"id":"W3176075285","doi":"10.1145/3462741.3466809","title":"Bigger Isn’t Better: The Ethical and Scientific Vices of Extra-Large Datasets in Language Models","year":2021,"lang":"en","type":"article","venue":"","topic":"Mobile Crowdsensing and Crowdsourcing","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"Social Sciences and Humanities Research Council of Canada; Mitacs; Compute Canada","keywords":"Computer science; Quality (philosophy); Data science; Language model; Control (management); Artificial intelligence; Epistemology","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.3646787,0.001353158,0.002259151,0.00397947,0.008335415,0.02060191,0.007789428,0.0140077,0.003825767],"category_scores_gemma":[0.6478156,0.00202591,0.00212364,0.005891038,0.05928051,0.04195077,0.01763282,0.02909746,0.002570196],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.007593858,"about_ca_system_score_gemma":0.009054673,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008412826,"about_ca_topic_score_gemma":0.009235639,"domain_scores_codex":[0.6336465,0.2773532,0.01103793,0.01898085,0.05723251,0.001748933],"domain_scores_gemma":[0.1873844,0.6770548,0.01283396,0.08643746,0.03266397,0.003625516],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000367796,0.0001248167,0.006198021,0.000630131,0.0002311562,0.0002478901,0.01432224,0.00264005,0.0008226775,0.7671307,0.145643,0.06164153],"study_design_scores_gemma":[0.0001616462,0.00007185269,0.001896923,0.001200884,0.00004951431,0.0003234428,0.003682703,0.007779424,0.001595511,0.8105891,0.1724622,0.0001867555],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.01116766,0.005577194,0.138346,0.8229407,0.004451709,0.0002323005,0.001014875,0.00054808,0.01572149],"genre_scores_gemma":[0.3474365,0.005577754,0.2569085,0.3513451,0.01788127,0.003243552,0.001717115,0.002724386,0.01316577],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6353214,"threshold_uncertainty_score":0.7834643,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01724107290708424,"score_gpt":0.26430836574723,"score_spread":0.2470672928401458,"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."}}