{"id":"W4415965747","doi":"10.2139/ssrn.5705186","title":"Open Technical Problems in Open-Weight AI Model Risk Management","year":2025,"lang":"","type":"preprint","venue":"SSRN Electronic Journal","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"Université de Montréal; McGill University; Mila - Quebec Artificial Intelligence Institute; University of Toronto; Institute on Governance","funders":"","keywords":"Key (lock); Risk management; Openness to experience; Risk assessment; Training (meteorology); Best practice","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","sts","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaepi_narrow","open_science"],"category_scores_codex":[0.02612295,0.001516612,0.002069333,0.001431673,0.002030473,0.007715598,0.0419204,0.001060603,0.00008532848],"category_scores_gemma":[0.0003607398,0.001538467,0.0005914463,0.00226415,0.0003067761,0.003328959,0.07897813,0.04180677,0.00007183963],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.01153241,"about_ca_system_score_gemma":0.0290973,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002016059,"about_ca_topic_score_gemma":0.006010389,"domain_scores_codex":[0.9810396,0.00229441,0.003038784,0.003563256,0.001664241,0.00839971],"domain_scores_gemma":[0.9931199,0.0002579405,0.002521252,0.003266944,0.0004153928,0.0004185819],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.000124951,0.0004222537,0.001079032,0.0000850162,0.0004422844,0.00004229194,0.0003390532,0.5435295,0.000001565492,0.3786284,0.0001667452,0.07513898],"study_design_scores_gemma":[0.002756956,0.0002209773,0.0002051999,0.001140201,0.0001958473,0.0002097481,0.0001430351,0.4835409,0.000001741188,0.5082285,0.002467347,0.0008895479],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0002995809,0.002433457,0.9475759,0.00505453,0.001701067,0.004778464,0.00001449326,0.0001105772,0.03803192],"genre_scores_gemma":[0.572614,0.09005703,0.2828697,0.001496221,0.0009401886,0.001106113,0.00003969185,0.0002727152,0.05060441],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6647062,"threshold_uncertainty_score":0.9997582,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01651247511547286,"score_gpt":0.313643105026537,"score_spread":0.2971306299110641,"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."}}