{"id":"W4401864111","doi":"10.23977/jaip.2024.070306","title":"The Training Process and Methods for LLMs Using an Own Knowledge Base","year":2024,"lang":"en","type":"article","venue":"Journal of Artificial Intelligence Practice","topic":"Digital Rights Management and Security","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Training (meteorology); Process (computing); Base (topology); Knowledge base; Computer science; Artificial intelligence; Geography; Mathematics","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.004998968,0.0001125736,0.0001494865,0.0001362875,0.0003291802,0.002302051,0.000605478,0.00003637678,0.000002614885],"category_scores_gemma":[0.001908902,0.00006923229,0.00008026696,0.0004048345,0.00009188161,0.005288528,0.00008357082,0.0002492012,0.000003739775],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003301437,"about_ca_system_score_gemma":0.0002335119,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003871163,"about_ca_topic_score_gemma":0.00001150231,"domain_scores_codex":[0.9986235,0.0002201282,0.0005025828,0.0002141111,0.0001965189,0.0002431057],"domain_scores_gemma":[0.9961992,0.002794348,0.0002755973,0.000173806,0.0004324758,0.0001245873],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004184718,0.00007677134,2.257593e-7,0.0000277645,0.00003559331,0.00001549724,0.007400494,0.000137917,0.0003631421,0.2547389,0.00002021914,0.7371417],"study_design_scores_gemma":[0.00001646118,0.0002920396,6.20663e-7,0.00006961433,0.00006338961,0.000126666,0.003859181,0.6824434,0.006958501,0.2252326,0.0808207,0.0001167744],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003662559,0.001454045,0.9899263,0.001956666,0.001319843,0.000149506,5.043074e-7,0.00003023573,0.001500332],"genre_scores_gemma":[0.5817627,0.0002092363,0.4170171,0.0001897722,0.0006522812,0.000005884219,2.581717e-7,0.00001807976,0.0001446805],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7370249,"threshold_uncertainty_score":0.9987336,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1852679543090059,"score_gpt":0.4814618281898133,"score_spread":0.2961938738808074,"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."}}