{"id":"W3099658661","doi":"10.18653/v1/2020.emnlp-main.744","title":"An information theoretic view on selecting linguistic probes","year":2020,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"ca_institutions":"Vector Institute; University of Toronto","funders":"Canadian Institute for Advanced Research","keywords":"Classifier (UML); Artificial intelligence; Modulo; Computer science; Information gain; Natural language processing; Selection (genetic algorithm); Machine learning; Mathematics; Discrete mathematics","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001415477,0.00007652288,0.00007013509,0.00004133076,0.00007197596,0.0002763772,0.0005523376,0.00003060652,0.00001190991],"category_scores_gemma":[0.0003981947,0.00005660379,0.00001534391,0.0003081389,0.00001230695,0.0007011463,0.00006736886,0.0001302554,0.00005304341],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001507222,"about_ca_system_score_gemma":0.00003111912,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005648021,"about_ca_topic_score_gemma":4.677194e-7,"domain_scores_codex":[0.9994322,0.00003066047,0.0001322259,0.0001369233,0.0001485804,0.0001194001],"domain_scores_gemma":[0.9995775,0.00003608916,0.00005473766,0.0001770687,0.00009200257,0.0000625631],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000003465563,0.00001144868,0.00002204825,0.00007179553,0.000001806654,0.000002623639,0.003688877,0.00001277212,0.0005615424,0.8474262,0.0001559228,0.1480415],"study_design_scores_gemma":[0.0002131897,0.001018531,0.0000371904,0.000169375,0.000008133288,0.00001848788,0.00007599274,0.551122,0.2023811,0.2399284,0.004502072,0.0005254675],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0009662933,0.0001193522,0.9894161,0.001389297,0.00004662002,0.0001366297,2.149578e-7,0.001685915,0.006239527],"genre_scores_gemma":[0.6704164,0.000001993487,0.32605,0.00347723,0.00004176213,0.000004449324,0.000001415146,0.000002591232,0.000004180527],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.66945,"threshold_uncertainty_score":0.2665112,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01090330771742997,"score_gpt":0.2687331560370854,"score_spread":0.2578298483196554,"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."}}