{"id":"W3152698349","doi":"10.18653/v1/2021.emnlp-main.230","title":"Masked Language Modeling and the Distributional Hypothesis: Order Word Matters Pre-training for Little","year":2021,"lang":"en","type":"article","venue":"Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing","topic":"Topic Modeling","field":"Computer Science","cited_by":182,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Word (group theory); Natural language processing; Artificial intelligence; Word order; Language model; Downstream (manufacturing); Order (exchange); Parametric statistics; Linguistics; Mathematics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003336436,0.001391239,0.0007572662,0.0004789316,0.000642718,0.001346939,0.001325684,0.00138786,0.00430266],"category_scores_gemma":[0.0177008,0.0006649107,0.0008272231,0.0005605142,0.0009945612,0.004688173,0.001508514,0.004304003,0.002496617],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000626173,"about_ca_system_score_gemma":0.00161977,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00397339,"about_ca_topic_score_gemma":0.008194104,"domain_scores_codex":[0.9984982,0.0006824813,0.00008020862,0.0005143941,0.0001157319,0.0001090177],"domain_scores_gemma":[0.9919951,0.005831776,0.0001936676,0.001395887,0.0003859006,0.0001977136],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001910816,0.000808427,0.01988021,0.0006366516,0.0003517652,0.0007194346,0.001343457,0.1790687,0.09119251,0.0205491,0.02816353,0.6553755],"study_design_scores_gemma":[0.00007076973,0.0003268118,0.003286589,0.00006710485,0.00006628449,0.0002239117,0.0002630102,0.9231259,0.03162676,0.03601005,0.004878645,0.00005419835],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3295034,0.001240609,0.6500712,0.002922499,0.0005073823,0.0001780202,0.001062519,0.008094657,0.006419796],"genre_scores_gemma":[0.8300682,0.0002766097,0.1596263,0.001347635,0.0001163681,0.0002293684,0.003277213,0.0007923447,0.004265942],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00430266,"threshold_uncertainty_score":0.01764494,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07636691541448742,"score_gpt":0.3862203416242639,"score_spread":0.3098534262097765,"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."}}