{"id":"W4287887656","doi":"10.18653/v1/2022.naacl-main.114","title":"Same Neurons, Different Languages: Probing Morphosyntax in Multilingual Pre-trained Models","year":2022,"lang":"en","type":"article","venue":"Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","topic":"Topic Modeling","field":"Computer Science","cited_by":17,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Vetenskapsrådet; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; National Science Foundation","keywords":"Computer science; Linguistics; Artificial intelligence; Natural language processing; 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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002071863,0.001313713,0.0007623113,0.000863812,0.0008403775,0.003555045,0.002071627,0.001535385,0.006097706],"category_scores_gemma":[0.01081723,0.001132556,0.001167208,0.001232653,0.001043754,0.005751746,0.003519366,0.00442494,0.002941853],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001430471,"about_ca_system_score_gemma":0.001198668,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01267828,"about_ca_topic_score_gemma":0.03495999,"domain_scores_codex":[0.9986669,0.0004900516,0.00007454786,0.0004745958,0.0001269355,0.000167092],"domain_scores_gemma":[0.9939851,0.004483548,0.0001199615,0.0007253057,0.0004916982,0.000194345],"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":[0.00355261,0.0006532403,0.0508715,0.0007972936,0.00124286,0.001538057,0.0061215,0.1563657,0.03830507,0.02001373,0.03229197,0.6882464],"study_design_scores_gemma":[0.0002029956,0.000236781,0.01521495,0.0001998042,0.0004696305,0.0007737657,0.002776214,0.8817226,0.02603533,0.05928672,0.0129149,0.0001663932],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7409058,0.003803312,0.2135179,0.002597322,0.0007781226,0.0001249054,0.00563301,0.008852161,0.02378748],"genre_scores_gemma":[0.9576008,0.0003787292,0.02997136,0.0002679758,0.00006909083,0.00008205803,0.005720489,0.001789743,0.004119745],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.01267828,"threshold_uncertainty_score":0.02520895,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01915628939806864,"score_gpt":0.2542868450759871,"score_spread":0.2351305556779185,"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."}}