{"id":"W7082254717","doi":"10.48448/5drz-7e13","title":"DRAMA: Diverse Augmentation from Large Language Models Towards Smaller Generalizable Dense Retrievers","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":"University of Waterloo","funders":"","keywords":"Inference; Robustness (evolution); Training set; Language model; Bridging (networking); Structured prediction","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":[0.0005399144,0.0003110883,0.0002918899,0.0003397295,0.0002993988,0.0003032676,0.002042944,0.0002687705,0.001200473],"category_scores_gemma":[0.0001510656,0.0002926518,0.00007477569,0.001057221,0.0002589343,0.0004476313,0.001104489,0.0002645421,0.0001083653],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001537567,"about_ca_system_score_gemma":0.0007848866,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002978749,"about_ca_topic_score_gemma":0.0002811647,"domain_scores_codex":[0.9973446,0.00005217879,0.0002609838,0.00104237,0.0007066631,0.0005932134],"domain_scores_gemma":[0.9984193,0.00003891801,0.0002233113,0.0009726313,0.0001868677,0.0001589881],"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":[0.00002431572,0.0004222898,0.0002542758,0.0001948893,0.000211469,0.0003658545,0.005025843,0.01153311,0.004738022,0.1206312,0.8276701,0.02892865],"study_design_scores_gemma":[0.001412658,0.00005513335,0.0000669233,0.0002830625,0.00006342045,0.000007512628,0.00127377,0.6944227,0.004988198,0.02186555,0.2744529,0.001108264],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"other","genre_scores_codex":[0.0001951219,0.0003518866,0.5182659,0.001012952,0.0009723108,0.0002261926,0.0001146692,0.0004035834,0.4784574],"genre_scores_gemma":[0.02760018,0.0001220411,0.1144593,0.00145827,0.000320912,0.00001675594,0.0002379918,0.0000327853,0.8557518],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.6828895,"threshold_uncertainty_score":0.9999526,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02177187000693468,"score_gpt":0.2651422261453536,"score_spread":0.2433703561384189,"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."}}