{"id":"W4405767395","doi":"10.48550/arxiv.2412.16971","title":"Part-Of-Speech Sensitivity of Routers in Mixture of Experts Models","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"Grand Équipement National De Calcul Intensif","keywords":"Security token; Computer science; Routing (electronic design automation); Natural language processing; Artificial intelligence; Computer network","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":[0.004947753,0.001014018,0.001271299,0.001152128,0.0006277493,0.002164932,0.001692596,0.002125392,0.00271038],"category_scores_gemma":[0.02486739,0.000944885,0.001034414,0.0006255786,0.001427055,0.004324565,0.00161037,0.00287771,0.0008073693],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001719625,"about_ca_system_score_gemma":0.0007537025,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008795805,"about_ca_topic_score_gemma":0.005634811,"domain_scores_codex":[0.9983974,0.0006740952,0.00004785376,0.0005082421,0.0001272202,0.0002451977],"domain_scores_gemma":[0.9852741,0.01161355,0.000890714,0.0008797267,0.0008242544,0.0005175258],"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":[0.001050324,0.000136528,0.01950878,0.0001172151,0.000249458,0.000356158,0.001232655,0.906675,0.006459489,0.03740111,0.001967432,0.02484578],"study_design_scores_gemma":[0.00001128666,0.00004246504,0.001434756,0.00001077532,0.00002679951,0.00006084774,0.00008722921,0.9828795,0.0008839724,0.01430725,0.0002336586,0.00002161155],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6702438,0.0005825958,0.3242372,0.0009112836,0.00006583144,0.00006764197,0.0002516114,0.0005536914,0.00308631],"genre_scores_gemma":[0.9804476,0.0001906263,0.01541755,0.000139632,0.00003335944,0.00005055129,0.0002968266,0.00009792691,0.003325859],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.008795805,"threshold_uncertainty_score":0.0261665,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08469193516844202,"score_gpt":0.1960231525919892,"score_spread":0.1113312174235471,"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."}}