{"id":"W4404782689","doi":"10.18653/v1/2024.emnlp-main.626","title":"ORPO: Monolithic Preference Optimization without Reference Model","year":2024,"lang":"en","type":"article","venue":"","topic":"Multi-Criteria Decision Making","field":"Decision Sciences","cited_by":76,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Institute for Information and Communications Technology Promotion; Ministry of Science and ICT, South Korea; Korea Advanced Institute of Science and Technology","keywords":"Preference; Computer science; Mathematics; Statistics","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":["scholarly_communication","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.002604226,0.0002557187,0.0003388679,0.0006048013,0.0001569808,0.001950968,0.001294732,0.0001801122,0.004156707],"category_scores_gemma":[0.00234742,0.0001699266,0.00009959618,0.00126972,0.00008647855,0.001225774,0.0003946185,0.0003171513,0.002670367],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007201155,"about_ca_system_score_gemma":0.0002495837,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002639313,"about_ca_topic_score_gemma":0.00003207672,"domain_scores_codex":[0.995257,0.0002000382,0.0009644027,0.001158763,0.002045783,0.0003739736],"domain_scores_gemma":[0.9968972,0.00104241,0.0001287086,0.00118151,0.0005605964,0.0001895703],"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":[0.00006849483,0.00004542328,0.0006782346,0.00001599742,0.00001680648,0.00001695517,0.0007398416,0.8349097,0.001167864,0.06145142,0.01417895,0.08671035],"study_design_scores_gemma":[0.0001221949,0.00001991827,0.0001730888,0.00006962578,0.000008304071,0.000009590251,0.0001001125,0.9347044,0.0002804012,0.05915632,0.005133229,0.0002227636],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02452672,0.0002004001,0.9124548,0.0005057315,0.0005364397,0.0002394434,0.00002169509,0.0004317782,0.06108294],"genre_scores_gemma":[0.7990939,0.00003050448,0.1773045,0.000261999,0.00005498368,0.00002417279,0.000006171858,0.00002588914,0.02319787],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7745672,"threshold_uncertainty_score":0.9990851,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5224362640997373,"score_gpt":0.47926384642419,"score_spread":0.0431724176755473,"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."}}