{"id":"W4412888901","doi":"10.18653/v1/2025.fever-1.19","title":"SANCTUARY: An Efficient Evidence-based Automated Fact Checking System","year":2025,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of New Brunswick","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science","routes":{"ca_aff":true,"ca_fund":true,"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":[],"consensus_categories":[],"category_scores_codex":[0.0004731637,0.0001341399,0.0001666437,0.0001724514,0.0001411569,0.0002452368,0.000909154,0.00006020003,0.00001125331],"category_scores_gemma":[0.00005324393,0.0001113029,0.00005438899,0.0005018179,0.00001528501,0.0003086653,0.0001712387,0.00008801822,0.00003782188],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001793776,"about_ca_system_score_gemma":0.0002250257,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001643601,"about_ca_topic_score_gemma":0.000006961794,"domain_scores_codex":[0.9985965,0.00009706522,0.0002674241,0.000488037,0.0002725258,0.0002784538],"domain_scores_gemma":[0.998778,0.0001277826,0.00005478286,0.0008793338,0.00008342903,0.00007665472],"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":[0.00001416725,0.0001447737,0.002976337,0.000434675,0.00003700369,0.00004850726,0.0007117091,0.6319367,0.003963178,0.310616,0.0009156067,0.04820134],"study_design_scores_gemma":[0.0001582941,0.00002155224,0.001895369,0.0004553587,0.000004788343,0.000001861662,0.00005269397,0.9928371,0.004333803,0.00003014,0.00009011538,0.0001189637],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1097448,0.0001345009,0.8824106,0.000574747,0.0005575393,0.0001826043,2.635107e-7,0.003065473,0.003329498],"genre_scores_gemma":[0.9407116,4.997802e-7,0.05863985,0.0004621103,0.00002376968,0.00001276436,5.738608e-7,0.000004548479,0.0001442499],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8309669,"threshold_uncertainty_score":0.45388,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04386299318024332,"score_gpt":0.2970264958567863,"score_spread":0.253163502676543,"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."}}