{"id":"W4385570658","doi":"10.18653/v1/2023.findings-acl.858","title":"Exploring the Effectiveness of Prompt Engineering for Legal Reasoning Tasks","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":35,"is_retracted":false,"has_abstract":true,"ca_institutions":"Thomson Reuters (Canada)","funders":"","keywords":"Task (project management); Computer science; Textual entailment; Natural language processing; Artificial intelligence; Cluster analysis; Logical consequence; Shot (pellet); Zero (linguistics); Best practice; Natural language; Linguistics; Engineering","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":[0.0009863005,0.00005965958,0.00008903568,0.00005698659,0.0000515171,0.00004381466,0.00040186,0.00001172765,4.646001e-7],"category_scores_gemma":[0.00015491,0.00004225138,0.00004381338,0.0002822508,0.000005821315,0.0003479288,0.000159989,0.00005157811,0.000003347083],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001481756,"about_ca_system_score_gemma":0.00001898466,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004342272,"about_ca_topic_score_gemma":0.000001158662,"domain_scores_codex":[0.9994079,0.00002938689,0.0001050225,0.0001720908,0.0001123715,0.0001732587],"domain_scores_gemma":[0.9990917,0.0005377098,0.00001989791,0.0002963299,0.00003423527,0.00002007864],"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":[0.000009902948,0.0000091753,0.0003127287,0.0002584502,0.00002752135,0.000003492207,0.0006585547,0.1339632,0.02007971,0.8207766,0.00002165228,0.02387902],"study_design_scores_gemma":[0.000130916,0.00002171387,0.005883256,0.00008431015,0.000002206943,0.000002299464,0.00002064026,0.9509904,0.04192558,0.0005057752,0.000363751,0.000069191],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3055908,0.00000676944,0.6934623,0.00007076405,0.0002777952,0.0001791189,2.472495e-7,0.0001882931,0.0002239392],"genre_scores_gemma":[0.9639861,0.000001474547,0.03572498,0.000005934795,0.00004719951,0.0001750989,4.280591e-7,0.000006867405,0.000051946],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8202708,"threshold_uncertainty_score":0.1722961,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06120192517994883,"score_gpt":0.2559378797659381,"score_spread":0.1947359545859893,"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."}}