{"id":"W4389519946","doi":"10.18653/v1/2023.findings-emnlp.300","title":"Verb Conjugation in Transformers Is Determined by Linear Encodings of Subject Number","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"York University; National Science Foundation","keywords":"Verb; Subject (documents); Computer science; Transformer; Encoding (memory); Mathematics Subject Classification; Position (finance); Artificial intelligence; Layer (electronics); Linguistics; Natural language processing; Arithmetic; Speech recognition; Mathematics; Discrete mathematics; Philosophy; Physics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009501242,0.0005701869,0.0003868247,0.0004584265,0.0003791458,0.003138139,0.0007772252,0.0005550539,0.008371241],"category_scores_gemma":[0.006633854,0.0006040847,0.0006467816,0.0004158708,0.001759114,0.007276867,0.002032232,0.001727799,0.002016515],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008935937,"about_ca_system_score_gemma":0.0007684987,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001366421,"about_ca_topic_score_gemma":0.001513598,"domain_scores_codex":[0.9993268,0.0001957098,0.00004451115,0.000233881,0.0001111951,0.00008794166],"domain_scores_gemma":[0.9978181,0.001060941,0.0002420591,0.0005773757,0.0001730435,0.0001284037],"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":[0.0007828156,0.0001277973,0.01014204,0.0003790731,0.00009037867,0.000437988,0.003171694,0.01842511,0.1447393,0.5374826,0.005235577,0.2789855],"study_design_scores_gemma":[0.00007119741,0.0002610386,0.00595467,0.00007770864,0.0001465048,0.0004090503,0.0007530929,0.2119991,0.07697494,0.6875709,0.01570393,0.00007787601],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2548369,0.0001224841,0.7012023,0.0008481873,0.0001410453,0.00007574331,0.0004593991,0.004455682,0.03785828],"genre_scores_gemma":[0.9518051,0.0001142811,0.04202811,0.0001468921,0.00002086925,0.00003799633,0.0003710427,0.0005928932,0.004882922],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008371241,"threshold_uncertainty_score":0.02800459,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01973994559083426,"score_gpt":0.269174178691716,"score_spread":0.2494342331008818,"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."}}