{"id":"W6948005058","doi":"10.48448/y38b-tb31","title":"What explains the success of cross-modal fine-tuning with ORCA?","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Cell Image Analysis Techniques","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Modalities; Variety (cybernetics); Transformer; Training (meteorology); Modality (human–computer interaction)","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":[0.0004896249,0.0002329977,0.0002181072,0.0002448262,0.00008779111,0.0003704863,0.001209915,0.0001629868,0.0001460749],"category_scores_gemma":[0.00008496297,0.0001441742,0.00008178588,0.0005491654,0.001789131,0.00002082268,0.0005760158,0.0001848432,0.00001987367],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002204037,"about_ca_system_score_gemma":0.0002939427,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001949335,"about_ca_topic_score_gemma":0.001390387,"domain_scores_codex":[0.9984385,0.00002583507,0.0002154897,0.0006270006,0.0004134649,0.0002796545],"domain_scores_gemma":[0.998549,0.0000110659,0.0001896528,0.001017975,0.0001738611,0.00005842779],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00005722504,0.0001861234,0.001364533,0.000333332,0.0004403516,0.00006441933,0.0002909231,0.000143083,0.5962904,0.00101526,0.3856042,0.01421011],"study_design_scores_gemma":[0.0003150875,0.0004226413,0.00005265541,0.0006464397,0.0002017864,0.00004751183,0.0005186119,0.001516101,0.493539,0.000380172,0.501623,0.0007369427],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.04704114,0.02346327,0.03100568,0.001650959,0.001011474,0.001681227,0.0000877588,0.0005021585,0.8935564],"genre_scores_gemma":[0.4750623,0.001215336,0.00461349,0.0003155766,0.0006159035,0.00003888281,0.0001585166,0.0003447103,0.5176352],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.4280212,"threshold_uncertainty_score":0.6592126,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01184453542725999,"score_gpt":0.3229822157583338,"score_spread":0.3111376803310738,"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."}}