{"id":"W4243640523","doi":"10.31219/osf.io/m6gcn","title":"Attention Mechanism, Transformers, BERT, and GPT: Tutorial and Survey","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":80,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Transformer; Encoder; Computer science; Generative grammar; Artificial intelligence; Speech recognition; Electrical engineering; Engineering; Voltage","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":[0.001082362,0.001427824,0.001024348,0.001983147,0.0004518123,0.00196178,0.001705855,0.001380426,0.007973807],"category_scores_gemma":[0.002680471,0.0008333094,0.0008787505,0.003451914,0.001792342,0.005870752,0.001384212,0.00274147,0.003041854],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001835001,"about_ca_system_score_gemma":0.00123359,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003118063,"about_ca_topic_score_gemma":0.001770858,"domain_scores_codex":[0.9995607,0.00008991022,0.00004238601,0.0001414826,0.0001152578,0.00005028742],"domain_scores_gemma":[0.9994129,0.0003606756,0.0000362765,0.00008820081,0.00007633973,0.00002568795],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00004936412,0.00007013555,0.0007854472,0.001239268,0.00005361702,0.0001991458,0.0003213011,0.009000665,0.002095185,0.5549942,0.01835197,0.4128397],"study_design_scores_gemma":[0.00001610686,0.000128391,0.001328657,0.000521211,0.0001124399,0.001528227,0.000134839,0.07159153,0.004213911,0.6698891,0.2504554,0.0000801402],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"review","genre_scores_codex":[0.008556869,0.2306855,0.6933733,0.002884079,0.001105374,0.0001648877,0.0006547378,0.002012331,0.06056293],"genre_scores_gemma":[0.2486167,0.3513713,0.3347901,0.002460073,0.004508108,0.000621163,0.001887248,0.001253374,0.05449197],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.007973807,"threshold_uncertainty_score":0.02667511,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05466022564196835,"score_gpt":0.2753320060763013,"score_spread":0.2206717804343329,"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."}}