{"id":"W4399785296","doi":"10.1016/j.ipm.2024.103809","title":"Are LLMs good at structured outputs? A benchmark for evaluating structured output capabilities in LLMs","year":2024,"lang":"en","type":"article","venue":"Information Processing & Management","topic":"Imbalanced Data Classification Techniques","field":"Computer Science","cited_by":60,"is_retracted":false,"has_abstract":false,"ca_institutions":"McGill University","funders":"National Natural Science Foundation of China","keywords":"Benchmark (surveying); Geography","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","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0008874912,0.0002643468,0.0002285317,0.0006159737,0.000251345,0.00132457,0.0008684581,0.0001026092,0.00001621562],"category_scores_gemma":[0.000128681,0.0002470134,0.00006971702,0.0007406953,0.00005473807,0.004447567,0.0004117508,0.0001638491,0.00002630985],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005676851,"about_ca_system_score_gemma":0.00008577896,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000008096347,"about_ca_topic_score_gemma":0.00001868702,"domain_scores_codex":[0.997656,0.00004570781,0.0008284398,0.0004399373,0.0006283591,0.0004015173],"domain_scores_gemma":[0.9985792,0.00006692833,0.0004648839,0.0006155752,0.0002149628,0.00005845878],"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":[0.00002415075,0.0000210811,0.0003846017,0.003164658,0.00003826324,0.000003794759,0.007656362,0.001220104,0.0001194305,0.08471923,0.004162819,0.8984855],"study_design_scores_gemma":[0.001224229,0.0001100832,0.03583704,0.001215452,0.0000514811,0.00001859442,0.002337473,0.784241,0.003976021,0.06989072,0.1001521,0.0009458198],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01026838,0.0003653287,0.9797121,0.001276237,0.0007277673,0.001936305,0.00008951528,0.001303712,0.004320611],"genre_scores_gemma":[0.8460637,0.00001103997,0.1516653,0.0006331027,0.00004166251,0.0006831901,0.0002238756,0.0000179055,0.0006602959],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8975397,"threshold_uncertainty_score":0.9999982,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02699792293706213,"score_gpt":0.3020599101299465,"score_spread":0.2750619871928844,"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."}}