{"id":"W4417273487","doi":"10.2139/ssrn.5835542","title":"Event-Driven Alpha from Large Language Models: A Theoretical Framework for Contrastive Financial Representation Learning","year":2025,"lang":"","type":"preprint","venue":"SSRN Electronic Journal","topic":"Stock Market Forecasting Methods","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"Canadian Imperial Bank of Commerce (Canada)","funders":"","keywords":"Interpretation (philosophy); Representation (politics); Financial market; Focus (optics); Alpha (finance); Event (particle physics); Modern portfolio theory; Downside risk","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.004485473,0.0008456337,0.001030779,0.001429997,0.0004876044,0.002089289,0.00246919,0.001574923,0.004232793],"category_scores_gemma":[0.02323278,0.0006825009,0.001086289,0.00123106,0.001535993,0.004906724,0.003003839,0.004178516,0.0008716318],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001031986,"about_ca_system_score_gemma":0.0009226754,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001266993,"about_ca_topic_score_gemma":0.001402512,"domain_scores_codex":[0.9987558,0.0006013481,0.00007339906,0.0002744568,0.0001988533,0.00009622359],"domain_scores_gemma":[0.9828068,0.01450444,0.0006259339,0.001101817,0.0006773777,0.0002834871],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0002865231,0.0002982173,0.003093936,0.0002610554,0.0001842067,0.0003427685,0.0006992267,0.3117372,0.005514644,0.4225794,0.004024044,0.2509789],"study_design_scores_gemma":[0.000009886426,0.00004017754,0.0001950014,0.00001695922,0.00001146008,0.00003856234,0.00001961931,0.8163821,0.0004908451,0.1823623,0.0004205235,0.0000125227],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01430114,0.0001866202,0.9835048,0.0005479528,0.00004031928,0.00003042174,0.0001493582,0.0002376209,0.00100164],"genre_scores_gemma":[0.7503623,0.0007811214,0.2401829,0.0007708207,0.0003958125,0.0004796873,0.001017538,0.0002647416,0.005745126],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004485473,"threshold_uncertainty_score":0.02372175,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04087777380982899,"score_gpt":0.4059147191940084,"score_spread":0.3650369453841794,"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."}}