{"id":"W4401042062","doi":"10.18653/v1/2024.semeval-1.265","title":"Edinburgh Clinical NLP at SemEval-2024 Task 2: Fine-tune your model unless you have access to GPT-4","year":2024,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Engineering and Physical Sciences Research Council; National Institute for Health and Care Research; Canadian Institute of Steel Construction; UK Research and Innovation; Nvidia; Accenture; Cisco Systems","keywords":"SemEval; Task (project management); Computer science; Natural language processing; Artificial intelligence; Speech recognition; Engineering","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001001616,0.0003191775,0.0003962606,0.0002300553,0.0001378816,0.0009634867,0.002416669,0.0001992972,0.0004246257],"category_scores_gemma":[0.000118492,0.0002651411,0.0002495881,0.0005367326,0.00004417266,0.0009290901,0.003178165,0.0004329507,0.0004036831],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000155843,"about_ca_system_score_gemma":0.0002604068,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000915262,"about_ca_topic_score_gemma":0.0001515574,"domain_scores_codex":[0.9963944,0.00006084086,0.0008088903,0.001414686,0.0006809596,0.0006402478],"domain_scores_gemma":[0.9978058,0.0002194986,0.00005838282,0.001354431,0.0001321277,0.0004297234],"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":[0.00003914773,0.0002176509,0.003510112,0.0001227692,0.0001691059,0.0003359923,0.002371401,0.1624989,0.0007873752,0.04698737,0.5585111,0.224449],"study_design_scores_gemma":[0.000201188,0.00005826339,0.0001051232,0.00008050996,0.00002107353,0.00001816092,0.00001780033,0.9661618,0.0003965051,0.006994885,0.02557952,0.0003652086],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04491892,0.0002495113,0.9165361,0.01360084,0.003891161,0.0003453818,0.00001346578,0.0006937133,0.01975094],"genre_scores_gemma":[0.7319653,0.0000504395,0.08533011,0.00384322,0.001414801,0.00007039199,0.000006647892,0.00005361245,0.1772655],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.831206,"threshold_uncertainty_score":0.9999801,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1398657191291666,"score_gpt":0.4029602116996834,"score_spread":0.2630944925705168,"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."}}