{"id":"W4394844693","doi":"10.1371/journal.pdig.0000489","title":"Biomedical text readability after hypernym substitution with fine-tuned large language models","year":2024,"lang":"en","type":"article","venue":"PLOS Digital Health","topic":"Text Readability and Simplification","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; University of Toronto","funders":"","keywords":"Readability; Computer science; Substitution (logic); Unified Medical Language System; Natural language processing; Text simplification; Automatic summarization; Artificial intelligence; Sentence; Comprehension; Biomedical text mining; Language model; Text mining","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":[],"consensus_categories":[],"category_scores_codex":[0.000332165,0.0001657623,0.0002060331,0.0001059529,0.0001043759,0.0004408233,0.0003438719,0.00007791587,0.00001654987],"category_scores_gemma":[0.00004680593,0.0001221843,0.00005971562,0.0005399858,0.0001142313,0.001726063,0.0001065569,0.0002070925,0.0001681993],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002266597,"about_ca_system_score_gemma":0.0004803689,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006525447,"about_ca_topic_score_gemma":0.00009165597,"domain_scores_codex":[0.9980803,0.00005516594,0.000323567,0.0006247238,0.0004835719,0.0004326392],"domain_scores_gemma":[0.9988393,0.0001229528,0.00004433303,0.0006739081,0.00005762503,0.0002618452],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003008859,0.004384396,0.005970475,0.003119842,0.0001337089,0.0002955072,0.01775353,0.0000763016,0.0003998879,0.3192532,0.00131755,0.6469947],"study_design_scores_gemma":[0.001764729,0.003102771,0.03283994,0.001914306,0.00005655324,0.0004237714,0.00110763,0.8333518,0.002294075,0.07830669,0.04246571,0.002372028],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.591002,0.0009287627,0.4007359,0.004804768,0.0001424793,0.0004386216,0.0001971538,0.0007406494,0.001009563],"genre_scores_gemma":[0.9973135,0.000008754238,0.002000186,0.0002800597,0.00007307357,0.00005389552,0.00009636773,0.00001271681,0.0001614415],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8332755,"threshold_uncertainty_score":0.498253,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01983503389309309,"score_gpt":0.2663777210944307,"score_spread":0.2465426872013376,"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."}}