{"id":"W4397048993","doi":"10.48550/arxiv.2405.09999","title":"Reward Centering","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Alliance de recherche numérique du Canada; DeepMind; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Psychology","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0001904645,0.0002906362,0.0002468593,0.000288456,0.0001087216,0.0003342621,0.002240355,0.0002253917,0.00003160903],"category_scores_gemma":[0.000025639,0.0003432923,0.0002207001,0.0004723359,0.0000596262,0.000199127,0.007792076,0.0009733264,0.0008129932],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002724389,"about_ca_system_score_gemma":0.0001513922,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004128029,"about_ca_topic_score_gemma":0.000003422777,"domain_scores_codex":[0.9982241,0.00006615894,0.0001917129,0.001025242,0.0001145246,0.0003781937],"domain_scores_gemma":[0.9982241,0.00004893744,0.0001466723,0.001365196,0.00007581094,0.0001392976],"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":[0.000003047587,0.00000802569,0.0003007896,0.0001291351,0.00006291679,0.0004329238,0.0001516349,0.8598413,0.00001123578,0.1382547,0.0005764198,0.0002277894],"study_design_scores_gemma":[0.0001239164,0.00003001141,0.0001118618,0.0002331493,0.00004277514,0.000005615753,0.00001993408,0.9744933,0.0000512772,0.02084294,0.003682197,0.0003630801],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01828875,0.00007057699,0.961003,0.0001804083,0.002211785,0.0001730912,0.000002820957,0.0007669141,0.01730263],"genre_scores_gemma":[0.9855635,0.0001205157,0.003222283,0.00008434789,0.0001027489,4.442133e-7,0.000007055619,0.00002545758,0.01087366],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9672747,"threshold_uncertainty_score":0.999965,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07759720201491045,"score_gpt":0.1875801099626226,"score_spread":0.1099829079477122,"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."}}