{"id":"W3017779903","doi":"10.18653/v1/2020.findings-emnlp.109","title":"Quantifying the Contextualization of Word Representations with Semantic Class Probing","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Microsoft (Canada)","funders":"","keywords":"Contextualization; Computer science; Natural language processing; Task (project management); Artificial intelligence; Context (archaeology); Inference; Word (group theory); Layer (electronics); Embedding; Class (philosophy); Linguistics","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":[],"consensus_categories":[],"category_scores_codex":[0.0002435585,0.0001462643,0.0002272484,0.00006764925,0.0000973199,0.0002008096,0.001030457,0.00007262683,0.000007419153],"category_scores_gemma":[0.00007593152,0.00009469466,0.0000626097,0.000289498,0.00004682854,0.0001620746,0.001042702,0.0002979485,0.000004817511],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002324223,"about_ca_system_score_gemma":0.0001778785,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002821742,"about_ca_topic_score_gemma":0.0001828117,"domain_scores_codex":[0.9985109,0.0001135114,0.0003803165,0.0005219,0.0003325949,0.0001407706],"domain_scores_gemma":[0.9983489,0.0001443119,0.0003095732,0.0009921369,0.0001688483,0.00003621275],"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":[0.00001105172,0.00004262939,0.003719246,0.0005307302,0.000154264,0.00001086529,0.01651529,0.1795509,0.0009321625,0.7872903,0.000393259,0.01084932],"study_design_scores_gemma":[0.0001346636,0.00001387536,0.0005434916,0.0002552261,0.00002399225,0.000005015933,0.0004861837,0.9925588,0.0009674401,0.004766794,0.00009483771,0.000149648],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.00806463,0.00006151553,0.9825975,0.005917001,0.0002193047,0.0005300929,0.000001000076,0.0001693745,0.002439563],"genre_scores_gemma":[0.9246711,0.00001038533,0.07485113,0.0002266929,0.00004653393,0.00003595724,0.000006377715,0.00001094188,0.0001409439],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9166064,"threshold_uncertainty_score":0.3861536,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1208758373613004,"score_gpt":0.3257294327598204,"score_spread":0.20485359539852,"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."}}