{"id":"W4399201246","doi":"10.29173/wclawr112","title":"Innocence Discovery Lab - Harnessing large language models to surface data buried in wrongful conviction case documents","year":2024,"lang":"en","type":"article","venue":"The Wrongful Conviction Law Review","topic":"Digital and Cyber Forensics","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Université Laval; Université du Québec à Montréal","funders":"","keywords":"Innocence; Conviction; Computer science; Political science; Law","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":true,"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":[0.007800082,0.0004350311,0.0004456668,0.003746949,0.001461778,0.003334692,0.001257364,0.0008345513,0.002456635],"category_scores_gemma":[0.02617414,0.0003714398,0.0005742974,0.002018957,0.001642078,0.003392752,0.003334892,0.001976049,0.001258959],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00165396,"about_ca_system_score_gemma":0.004363753,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01472283,"about_ca_topic_score_gemma":0.02928747,"domain_scores_codex":[0.994929,0.002447122,0.0002871943,0.0007655903,0.001376752,0.0001944014],"domain_scores_gemma":[0.9804543,0.01221193,0.001333852,0.003754766,0.001904037,0.0003411547],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003266462,0.0007124341,0.07843784,0.000668704,0.0002150617,0.001762756,0.01018127,0.03569074,0.009867257,0.09393738,0.07371426,0.6944857],"study_design_scores_gemma":[0.000103987,0.0003014592,0.03581549,0.000725509,0.0001209933,0.001681188,0.01074392,0.4874195,0.04096419,0.1495697,0.2722593,0.0002946785],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3958624,0.002270274,0.4968555,0.0209591,0.0003522717,0.001413285,0.01922961,0.008187184,0.05487036],"genre_scores_gemma":[0.6193981,0.001199816,0.3472189,0.001067015,0.0001369785,0.0007909504,0.0162836,0.000471982,0.01343272],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01472283,"threshold_uncertainty_score":0.0412513,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03983930074807804,"score_gpt":0.3085035578465523,"score_spread":0.2686642570984742,"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."}}