{"id":"W4389524452","doi":"10.18653/v1/2023.emnlp-main.12","title":"Sparse Universal Transformer","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Transformer; Computer science; Inference; Computation; Scaling; Language model; Generalization; Theoretical computer science; Artificial intelligence; Algorithm; Computer engineering; 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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00006723765,0.00003107229,0.00003244518,0.00004960114,0.00002784145,0.00002354548,0.0002804287,0.00001516354,0.00005205187],"category_scores_gemma":[0.000002076293,0.00002693309,0.00001987143,0.0002459541,0.000005668643,0.0001965485,0.00002810078,0.00002937512,0.0006241637],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000007824743,"about_ca_system_score_gemma":0.0000183064,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002663672,"about_ca_topic_score_gemma":0.000008977604,"domain_scores_codex":[0.9996193,0.000005644874,0.00004620543,0.0001203511,0.00008592271,0.0001225836],"domain_scores_gemma":[0.9997619,0.00001296081,0.000004428809,0.000179911,0.00000863169,0.00003218988],"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":[9.538736e-7,0.00000916963,0.0003067918,0.000004494452,0.000006912801,0.00004083322,0.001217567,0.001650768,0.001133824,0.8154744,0.006062298,0.174092],"study_design_scores_gemma":[0.0001623464,0.00001636857,0.0008067864,0.000003416562,0.000001419677,0.000004403921,0.0001027323,0.947628,0.002767965,0.008186013,0.04019546,0.0001251353],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02234118,0.000002641485,0.9238781,0.002081805,0.0001707835,0.0000293565,1.316596e-7,0.0004657092,0.05103032],"genre_scores_gemma":[0.9590513,0.00000723833,0.03005278,0.0002781983,0.00003098127,0.000001279821,4.75732e-7,0.000002990393,0.01057475],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9459772,"threshold_uncertainty_score":0.8022569,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03619836200111303,"score_gpt":0.2409397845633939,"score_spread":0.2047414225622809,"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."}}