{"id":"W4389979427","doi":"10.4337/cilj.2023.02.08","title":"Large language models and the treaty interpretation game","year":2023,"lang":"en","type":"article","venue":"Cambridge International Law Journal","topic":"Law, AI, and Intellectual Property","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Treaty; Interpretation (philosophy); Analogy; International law; Political science; Argument (complex analysis); Narrative; Law; Law and economics; Sociology; Epistemology; Computer science; Linguistics","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":[0.0007697127,0.0001166381,0.0001349899,0.0001008587,0.0002733163,0.0005958158,0.0008208694,0.00003986775,0.00004120041],"category_scores_gemma":[0.0001384866,0.00006998058,0.0001104258,0.0001640332,0.0001697353,0.001028767,0.0002810441,0.0002553681,0.0001260799],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000604299,"about_ca_system_score_gemma":0.00003120832,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005653673,"about_ca_topic_score_gemma":0.00002564114,"domain_scores_codex":[0.9987786,0.000123195,0.0002732522,0.0002063675,0.0003889611,0.0002296424],"domain_scores_gemma":[0.9992246,0.0002055975,0.0001045928,0.0001975533,0.0001776433,0.00009003613],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00008266159,0.00001801252,0.00001345196,0.000003089036,0.00007268418,0.00005417893,0.004747159,0.0002508835,0.0001674477,0.9768367,0.007272787,0.01048097],"study_design_scores_gemma":[0.001354074,0.00004012907,0.0001593591,0.00003490485,0.000008975931,0.0005379729,0.0003401971,0.9608828,0.0003746869,0.01921969,0.01691897,0.0001282743],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03604809,0.0005621454,0.7962781,0.009734537,0.002177878,0.0001700404,0.00002489458,0.0002085647,0.1547958],"genre_scores_gemma":[0.993232,0.0002363552,0.0003033516,0.001557891,0.0004358174,0.000008366047,0.000009062359,0.00001039064,0.004206775],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9606319,"threshold_uncertainty_score":0.5745466,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0172333229506089,"score_gpt":0.2573812544151795,"score_spread":0.2401479314645706,"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."}}