{"id":"W2327588828","doi":"10.1515/cllt.2011.008","title":"Safe harbour: Ethics and accessibility in sociolinguistic corpus building","year":2011,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Globe; Corpus linguistics; Linguistics; Computer science; Applied linguistics; Computational linguistics; Sociology; Data sharing; Natural language processing; Psychology; Philosophy","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.003722935,0.0004146052,0.0005079851,0.0002870772,0.0004401224,0.0002769927,0.001002565,0.000376092,0.00001313315],"category_scores_gemma":[0.06953516,0.0003923326,0.00006177234,0.0002759593,0.0005526629,0.00005200259,0.0008830013,0.001544759,0.000003557308],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009580346,"about_ca_system_score_gemma":0.0002997161,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002100445,"about_ca_topic_score_gemma":0.0000356249,"domain_scores_codex":[0.9970224,0.0004096489,0.0006796938,0.0009617337,0.0003228485,0.000603707],"domain_scores_gemma":[0.9955117,0.002400426,0.0003424094,0.0007827642,0.0006833333,0.0002794438],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00005236905,0.00007320072,0.00393391,0.0002031851,0.00001460854,0.0002585934,0.007835351,3.838753e-7,0.0001173063,0.9769413,0.000006962819,0.01056286],"study_design_scores_gemma":[0.0004029805,0.0001059785,0.002698604,0.0003277442,0.00004245555,0.000038471,0.00005596655,0.003581907,0.001158856,0.9887118,0.002330816,0.0005444175],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0440216,0.04465536,0.8581573,0.0001246368,0.007399999,0.001066716,0.0000475862,0.00184211,0.04268473],"genre_scores_gemma":[0.7360911,0.0002185214,0.2627535,0.0003325642,0.000483994,0.00001259571,0.000002414821,0.00003078156,0.00007451147],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6920695,"threshold_uncertainty_score":0.9998528,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03852763765365219,"score_gpt":0.3088898463052056,"score_spread":0.2703622086515534,"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."}}