{"id":"W4393903283","doi":"10.48550/arxiv.2404.00727","title":"A Controlled Reevaluation of Coreference Resolution Models","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"McGill University; Canadian Institute for Advanced Research; Nvidia","keywords":"Coreference; Resolution (logic); Computer science; Natural language processing; Artificial intelligence","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.009315697,0.001512153,0.00108387,0.0007961949,0.0006291604,0.00208945,0.003389084,0.001486642,0.002266327],"category_scores_gemma":[0.04239606,0.0007621203,0.0008583929,0.0005954989,0.001146576,0.003998397,0.002091597,0.003561866,0.00106837],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001520381,"about_ca_system_score_gemma":0.001838901,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005656952,"about_ca_topic_score_gemma":0.00855287,"domain_scores_codex":[0.9955645,0.001984207,0.0002375131,0.001482467,0.0005118698,0.0002194985],"domain_scores_gemma":[0.9756153,0.01684503,0.0008018807,0.004797726,0.001639826,0.0003001454],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002541739,0.001367875,0.01496592,0.00140678,0.0008165718,0.0003161709,0.001089605,0.443616,0.06192978,0.01335524,0.009915293,0.448679],"study_design_scores_gemma":[0.0002402204,0.0007497368,0.002357848,0.00008859651,0.0002335091,0.0001560915,0.0001614202,0.9501151,0.03324652,0.008158772,0.004412307,0.00007981534],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6478376,0.003319191,0.3268587,0.001519104,0.0004791925,0.0004936812,0.001102765,0.008898792,0.009490908],"genre_scores_gemma":[0.8947882,0.0004983156,0.09945504,0.0004945656,0.0000643981,0.0004518649,0.00106334,0.0005635443,0.002620748],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.009315697,"threshold_uncertainty_score":0.0492667,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1128739914334008,"score_gpt":0.2281634177091394,"score_spread":0.1152894262757386,"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."}}