{"id":"W4417287532","doi":"10.2139/ssrn.5912869","title":"LLMs as Judges: Toward The Automatic Review of GSN-compliant Assurance Cases","year":2025,"lang":"","type":"preprint","venue":"SSRN Electronic Journal","topic":"Safety Systems Engineering in Autonomy","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"York University; University of Ottawa","funders":"","keywords":"Leverage (statistics); Quality assurance; Harm; Task (project management); Obstacle; Risk assessment","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","research_integrity"],"consensus_categories":["metaepi_narrow"],"category_scores_codex":[0.007571647,0.001447375,0.002623372,0.0004716547,0.0002904047,0.0001603631,0.003211472,0.0005844246,0.0003646149],"category_scores_gemma":[0.001620315,0.001193878,0.001534361,0.0009618,0.0001805962,0.0002361122,0.0005510558,0.01111464,0.0002202394],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.006503644,"about_ca_system_score_gemma":0.01320741,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003189182,"about_ca_topic_score_gemma":0.000101031,"domain_scores_codex":[0.9888407,0.0006166415,0.003728103,0.0008366511,0.001230393,0.004747539],"domain_scores_gemma":[0.9940623,0.001377986,0.001612938,0.002148987,0.0005536556,0.0002441256],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.0001030839,0.0004770431,0.0005586565,0.3483664,0.0222469,0.0002917243,0.003992134,0.3430841,0.0004227885,0.117727,0.006021026,0.1567091],"study_design_scores_gemma":[0.003905216,0.001507843,0.003121683,0.6212042,0.006887232,0.04464867,0.006374731,0.1223993,0.0007566533,0.06801274,0.114247,0.006934723],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.002220125,0.8751681,0.09905612,0.005973471,0.006572925,0.003065253,0.0000981875,0.0004641693,0.00738164],"genre_scores_gemma":[0.4899417,0.5066596,0.00064531,0.0002267851,0.0006780291,0.0002070203,0.00001814814,0.0001604348,0.001463043],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.4877215,"threshold_uncertainty_score":0.9998276,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01453619147121232,"score_gpt":0.2614246803901264,"score_spread":0.2468884889189141,"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."}}