{"id":"W4399452441","doi":"10.2139/ssrn.4851778","title":"LOLA: LLM-Assisted Online Learning Algorithm for Content Experiments","year":2024,"lang":"en","type":"preprint","venue":"SSRN Electronic Journal","topic":"Advanced Bandit Algorithms Research","field":"Decision Sciences","cited_by":12,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Content (measure theory); Online learning; Algorithm; Artificial intelligence; Multimedia; Mathematics","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","scholarly_communication","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.01124804,0.000677076,0.001112378,0.001166651,0.0006402345,0.001350051,0.002621009,0.000518168,0.0001406614],"category_scores_gemma":[0.00302411,0.0005118955,0.001013744,0.0007263587,0.0001899926,0.0002621759,0.002216902,0.01496461,0.0002860252],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.004227716,"about_ca_system_score_gemma":0.006825887,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005646952,"about_ca_topic_score_gemma":0.0001763067,"domain_scores_codex":[0.9876735,0.0005767925,0.001801504,0.001550308,0.003696075,0.004701797],"domain_scores_gemma":[0.994971,0.001129071,0.0009707944,0.0008353951,0.001707108,0.0003866582],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001172738,0.0002421997,0.00003391056,0.00002006623,0.0006557786,0.00003880047,0.0002651927,0.001757111,0.0003076573,0.001069045,0.001151244,0.9943417],"study_design_scores_gemma":[0.002078363,0.001391565,0.0001433666,0.0002690506,0.0001236428,0.0006458485,0.01124093,0.09869276,0.0005509637,0.8490093,0.03499835,0.0008558173],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0126134,0.02815709,0.9519025,0.001748614,0.00376095,0.001184381,0.0002389862,0.0001686511,0.0002254428],"genre_scores_gemma":[0.5381346,0.02357855,0.09526119,0.0003504351,0.01634931,0.0009025557,0.0007491248,0.0009694368,0.3237048],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9934859,"threshold_uncertainty_score":0.9997333,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1884793554634404,"score_gpt":0.4652270319985441,"score_spread":0.2767476765351037,"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."}}