{"id":"W6929451287","doi":"10.48448/fccs-5q95","title":"FLIRT: Feedback Loop In-context Red Teaming","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Lipid Membrane Structure and Behavior","field":"Biochemistry, Genetics and Molecular Biology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Adversarial system; Robustness (evolution); Feedback loop; Generative grammar; Human-in-the-loop; Key (lock)","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002968144,0.001261889,0.0008141122,0.0005445629,0.000698688,0.001131497,0.003082553,0.00217916,0.008747319],"category_scores_gemma":[0.01152157,0.0005474107,0.0008086922,0.000135297,0.001622968,0.002251271,0.004211633,0.00242086,0.002183346],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009615422,"about_ca_system_score_gemma":0.001071868,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002276707,"about_ca_topic_score_gemma":0.002434606,"domain_scores_codex":[0.9983553,0.0005340195,0.00006064935,0.0004192982,0.0004484044,0.0001821857],"domain_scores_gemma":[0.9967529,0.001520338,0.0003394867,0.0007174208,0.0003943596,0.0002754073],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0007200143,0.0004536353,0.004822135,0.0002500466,0.0001170122,0.0007996738,0.0006874757,0.7045125,0.02867598,0.02854279,0.01685685,0.213562],"study_design_scores_gemma":[0.0000178321,0.00006676884,0.00009752152,0.000009783401,0.000006389195,0.00003913278,0.00001971245,0.9855382,0.004996591,0.007832548,0.001361974,0.00001350633],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"other","genre_scores_codex":[0.03482335,0.0001263419,0.9322463,0.000466022,0.0001906096,0.0001820472,0.0001097832,0.0271739,0.004681583],"genre_scores_gemma":[0.8060556,0.0000744489,0.1827532,0.0006782109,0.00006652701,0.0002020062,0.0002977249,0.001760355,0.008111933],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.008747319,"threshold_uncertainty_score":0.02926266,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01160376732455667,"score_gpt":0.2764850778676766,"score_spread":0.2648813105431199,"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."}}