{"id":"W4400406053","doi":"10.2139/ssrn.4856254","title":"Filtered not Mixed: Stochastic Filtering-Based Online Gating for Mixture of Large Language Models","year":2024,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Dalhousie University; Vector Institute; McMaster University; University of Toronto","funders":"","keywords":"Computer science; Gating; Language model; Artificial intelligence; Natural language processing; Psychology; Neuroscience","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":["metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.001871065,0.0004222073,0.0005819431,0.0003385244,0.0001300499,0.0002113046,0.001810769,0.0002985574,0.000006081485],"category_scores_gemma":[0.000136717,0.0003948719,0.0004812017,0.0001719345,0.00002315322,0.0001451095,0.001106757,0.004295186,0.000002790104],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006751743,"about_ca_system_score_gemma":0.003795932,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005260268,"about_ca_topic_score_gemma":0.0005418382,"domain_scores_codex":[0.9956548,0.0001012441,0.0007944662,0.0007279706,0.0005223326,0.002199247],"domain_scores_gemma":[0.9980986,0.0001771161,0.0005505476,0.0008457744,0.0002132456,0.000114724],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00005837654,0.0001896301,0.000003043081,0.0008208278,0.0004239123,0.00002116359,0.001721589,0.697566,0.002128701,0.2606289,0.00006845102,0.03636936],"study_design_scores_gemma":[0.0005211402,0.0001252631,0.000001619549,0.0004975134,0.00005989032,0.0000457974,0.0002062522,0.7473589,0.0006781606,0.2502182,0.00001532645,0.0002719249],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04374983,0.00440194,0.9488956,0.00102872,0.001128763,0.000411605,0.0002216351,0.0001370423,0.00002489496],"genre_scores_gemma":[0.9192816,0.00005387006,0.07940243,0.0001585333,0.0006694531,0.00003493633,0.00005965489,0.00006338669,0.0002761262],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8755318,"threshold_uncertainty_score":0.9998503,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02135863256316978,"score_gpt":0.2869592703647592,"score_spread":0.2656006378015894,"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."}}