{"id":"W4416536103","doi":"10.1038/s41597-025-05513-8","title":"Twenty years (2000–2020) of butterfly monitoring data across the contiguous United States","year":2025,"lang":"en","type":"article","venue":"Scientific Data","topic":"Lepidoptera: Biology and Taxonomy","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Cascades (Canada)","funders":"U.S. Geological Survey; U.S. Fish and Wildlife Service; Israel Institute for Biological Research; U.S. Department of the Interior; National Science Foundation","keywords":"Butterfly; Taxonomy (biology); Biodiversity; Population; Taxonomic rank; Occupancy","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0009521024,0.0005209173,0.0003756898,0.002992801,0.0003889672,0.0006470251,0.0007625282,0.0004574524,0.003540569],"category_scores_gemma":[0.002941287,0.0002933614,0.0004253908,0.005993539,0.0001483653,0.0006058729,0.001082161,0.000543372,0.001767695],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000885959,"about_ca_system_score_gemma":0.001904435,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.1501305,"about_ca_topic_score_gemma":0.1474841,"domain_scores_codex":[0.9993526,0.00008006562,0.00009776185,0.0001794316,0.0002023178,0.00008773919],"domain_scores_gemma":[0.9973789,0.0002559565,0.000824355,0.0002192705,0.001015666,0.0003059392],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002709255,0.0001464271,0.5118684,0.0008957489,0.0005318836,0.0002813593,0.0003785773,0.004091043,0.001026511,0.00111965,0.4137384,0.06565116],"study_design_scores_gemma":[0.00005811236,0.00006720886,0.7928913,0.0004058464,0.0001630809,0.0002156881,0.0004993753,0.003097493,0.0006633499,0.0006688843,0.20123,0.00003980145],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.09780091,0.001660901,0.00138047,0.0009930956,0.0001427131,0.00005880542,0.8922485,0.0003867063,0.005327955],"genre_scores_gemma":[0.09036713,0.0010908,0.002744132,0.000505454,0.00007421606,0.0002379561,0.9030937,0.00004617233,0.001840392],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.1501305,"threshold_uncertainty_score":0.2985134,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04054657016238197,"score_gpt":0.3212964209576983,"score_spread":0.2807498507953163,"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."}}