{"id":"W7082254231","doi":"10.48448/cdtf-hm31","title":"Robust Utility-Preserving Text Anonymization Based on Large Language Models","year":2025,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Queen's University","funders":"","keywords":"Robustness (evolution); Context (archaeology); Key (lock); Data modeling; Information privacy; Language model; Code (set theory); Face (sociological concept)","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":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001003132,0.0003028605,0.0002681945,0.0005666674,0.0003119331,0.0002787043,0.00295055,0.0002418288,0.001065823],"category_scores_gemma":[0.0004782245,0.0002765093,0.00006321857,0.001512667,0.0002144154,0.0003137869,0.0009024704,0.0003301603,0.00006080476],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007173925,"about_ca_system_score_gemma":0.0007588061,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001503548,"about_ca_topic_score_gemma":0.0001043683,"domain_scores_codex":[0.9972747,0.00006943108,0.0002546059,0.001088312,0.0007109466,0.0006019995],"domain_scores_gemma":[0.997776,0.0001088859,0.0001980536,0.001600235,0.0001875216,0.0001293065],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001692919,0.0008809619,0.0003119642,0.0007499399,0.0000445064,0.0001168817,0.00077407,0.3173162,0.0007191592,0.2114777,0.4261032,0.04148841],"study_design_scores_gemma":[0.0002163148,0.00002105179,0.00001434823,0.0002304543,0.000005543047,0.000001129392,0.00005220048,0.9405611,0.0003173372,0.00287616,0.05543873,0.0002656736],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.000002032711,0.00006876712,0.4939845,0.00103202,0.0002223163,0.0001386191,0.00001772728,0.0002823544,0.5042517],"genre_scores_gemma":[0.1096395,0.00001937819,0.1000769,0.002473003,0.00026594,0.00002758015,0.0001153351,0.00004559557,0.7873368],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.6232448,"threshold_uncertainty_score":0.9999687,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02542903830547049,"score_gpt":0.2561528963576901,"score_spread":0.2307238580522196,"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."}}