{"id":"W4389518966","doi":"10.18653/v1/2023.findings-emnlp.285","title":"Measuring the Knowledge Acquisition-Utilization Gap in Pretrained Language Models","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal; Mila - Quebec Artificial Intelligence Institute; Canadian Institute for Advanced Research; McGill University","funders":"","keywords":"Computer science; USable; Domain knowledge; Robustness (evolution); Parametric statistics; Task (project management); Measure (data warehouse); Knowledge acquisition; Artificial intelligence; Knowledge extraction; Machine learning; Data mining; Engineering; Mathematics","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":[0.0004775729,0.00006163268,0.00006023658,0.0001138457,0.0000647391,0.00007278409,0.0004738991,0.00003156196,0.00001541016],"category_scores_gemma":[0.00002132842,0.00004486095,0.00002292896,0.0006727552,0.000009162529,0.0003419149,0.000179077,0.00006660226,0.00007920091],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000326516,"about_ca_system_score_gemma":0.00003217142,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003891169,"about_ca_topic_score_gemma":0.00008875973,"domain_scores_codex":[0.9992697,0.00005987849,0.0001358805,0.0002067405,0.0001501929,0.0001775726],"domain_scores_gemma":[0.9994826,0.00007122607,0.00001895997,0.0003704133,0.00003193866,0.00002481669],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000003608934,0.00005743101,0.0007059943,0.00004869748,0.00001188463,0.00002175972,0.05794004,0.2463123,0.0021251,0.6075303,0.0007544013,0.08448844],"study_design_scores_gemma":[0.000110016,0.000003463626,0.001107357,0.00001523464,6.644445e-7,0.000001194917,0.0002784022,0.9848591,0.0004421843,0.01309926,0.00002538215,0.00005773098],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05710322,0.0001201513,0.9172422,0.0006691391,0.0001391414,0.0001358172,3.308933e-7,0.0003672193,0.02422282],"genre_scores_gemma":[0.9950644,0.000007133272,0.003673694,0.0001209192,0.00004525152,0.00001757587,0.000001945807,0.000005183864,0.001063861],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9379612,"threshold_uncertainty_score":0.1829376,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1207953153189931,"score_gpt":0.3006593014070115,"score_spread":0.1798639860880183,"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."}}