{"id":"W6948019733","doi":"10.48448/9cnb-s768","title":"Semantically-Prompted Language Models Improve Visual Descriptions","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Subterranean biodiversity and taxonomy","field":"Earth and Planetary Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Visual language; Key (lock); Semantics (computer science); Visualization; Language model; Image (mathematics); Visual reasoning","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":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0003343159,0.0002861854,0.0002568333,0.0006537241,0.0002068689,0.0003259063,0.0006967718,0.0002002948,0.01492879],"category_scores_gemma":[0.00002914326,0.0002292877,0.00008657899,0.0006751059,0.0009385484,0.0003200483,0.00008828863,0.0003465149,0.008180305],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001816035,"about_ca_system_score_gemma":0.0002697273,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004339939,"about_ca_topic_score_gemma":0.005085208,"domain_scores_codex":[0.9979289,0.00002783681,0.0001955882,0.0007396313,0.0005700887,0.000537973],"domain_scores_gemma":[0.999247,0.00003208134,0.00008742168,0.0003325772,0.00004046616,0.0002604666],"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":[0.0000739758,0.0003501394,0.005626953,0.0007403171,0.0004302358,0.001016316,0.002704433,0.0008615296,0.0009015218,0.004818405,0.7417744,0.2407017],"study_design_scores_gemma":[0.0009316138,0.0005811986,0.001014594,0.0005743933,0.0004259957,0.0001099714,0.003635659,0.5901919,0.0002645224,0.006245793,0.3936097,0.002414606],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.001767919,0.00156405,0.002830302,0.0002541989,0.002210882,0.0004505294,0.0007552279,0.0005956444,0.9895713],"genre_scores_gemma":[0.4158416,0.0001246352,0.01569101,0.0008863758,0.001190152,0.000003614387,0.0006873556,0.0001055535,0.5654697],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.5893304,"threshold_uncertainty_score":0.9925919,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03441052405999891,"score_gpt":0.2456936735914689,"score_spread":0.21128314953147,"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."}}