{"id":"W4411950007","doi":"10.1109/tit.2025.3584013","title":"Next-Token Prediction Capacity: General Upper Bounds and a Lower Bound for Transformers","year":2025,"lang":"en","type":"article","venue":"IEEE Transactions on Information Theory","topic":"Computational Physics and Python Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Upper and lower bounds; Computer science; Security token; Capacity planning; Mathematics; Algorithm; Computer network","routes":{"ca_aff":true,"ca_fund":true,"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":[0.007186194,0.003080759,0.003166832,0.002532593,0.002121809,0.004908658,0.006610891,0.003538772,0.0170725],"category_scores_gemma":[0.07216535,0.001818256,0.002818879,0.001952053,0.005991324,0.01880323,0.009259554,0.009474492,0.003117629],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004368857,"about_ca_system_score_gemma":0.003381656,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003409901,"about_ca_topic_score_gemma":0.002926412,"domain_scores_codex":[0.995636,0.0008650346,0.0002634216,0.001205826,0.000880682,0.001149008],"domain_scores_gemma":[0.9253954,0.06169772,0.002033064,0.006360305,0.002483673,0.002029886],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.001194543,0.0003830702,0.003984415,0.001228423,0.0001992117,0.0006709742,0.000937436,0.3251007,0.01330632,0.5482484,0.0140043,0.09074226],"study_design_scores_gemma":[0.00002655777,0.0001014883,0.000472014,0.0001985995,0.0000778535,0.0003559431,0.00007586189,0.6215416,0.006312638,0.368576,0.002193059,0.00006838886],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05017166,0.004837918,0.9109337,0.004167548,0.0003318847,0.000175161,0.001728634,0.002509917,0.02514358],"genre_scores_gemma":[0.8647054,0.004632843,0.1120141,0.00165123,0.0008218754,0.0007561942,0.001804714,0.001849624,0.01176397],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.0170725,"threshold_uncertainty_score":0.05711317,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01034835890367836,"score_gpt":0.2351723872306748,"score_spread":0.2248240283269964,"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."}}