{"id":"W4411120250","doi":"10.18653/v1/2025.wnu-1.7","title":"A Theoretical Framework for Evaluating Narrative Surprise in Large Language Models","year":2025,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Surprise; Narrative; Computer science; Natural language processing; Linguistics; Psychology; Philosophy; Communication","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0008814068,0.00008191536,0.0001327649,0.00008605888,0.00006520973,0.00006810598,0.000439722,0.00007634899,0.00004877139],"category_scores_gemma":[0.0004467932,0.00006970233,0.00004431691,0.0002591705,0.00002100371,0.0001883095,0.0002113996,0.0001485039,0.000002492314],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003878871,"about_ca_system_score_gemma":0.00008858065,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001242865,"about_ca_topic_score_gemma":0.00001994954,"domain_scores_codex":[0.9990087,0.00008160841,0.0001952838,0.0003143502,0.0001396812,0.0002604421],"domain_scores_gemma":[0.9989823,0.000529383,0.00002406332,0.0003796785,0.00005470603,0.00002990688],"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":[0.000005300218,0.00002436276,0.00005927423,0.000008641717,0.00000377266,0.00000156573,0.01196472,0.001716825,0.00002399341,0.9790666,0.00002340937,0.007101561],"study_design_scores_gemma":[0.0001670657,0.000008849119,0.00001283965,0.00003102142,9.8706e-7,1.61553e-7,0.0008387287,0.5225697,0.0001013856,0.4762259,0.000002914629,0.00004044545],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04167064,0.00009396383,0.9486452,0.001463284,0.0001356421,0.0002922381,0.000001038849,0.00008800464,0.007609939],"genre_scores_gemma":[0.585418,5.590418e-7,0.4138225,0.0005498728,0.00001195525,0.00004000325,4.024715e-7,0.000002374597,0.0001542993],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5437474,"threshold_uncertainty_score":0.2842378,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02994904019277738,"score_gpt":0.3667034877914282,"score_spread":0.3367544475986508,"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."}}