{"id":"W7125641124","doi":"10.1109/cascon66301.2025.00040","title":"Evaluating Generative AI for CS1 Code Grading: Direct Vs Reverse Methods","year":2025,"lang":"","type":"article","venue":"","topic":"Teaching and Learning Programming","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of British Columbia, Okanagan Campus; University of British Columbia; Okanagan University College; Okanagan Science Centre","funders":"","keywords":"Code (set theory); Source code; Generative grammar; Feature (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":["metaepi_narrow","sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.01274575,0.0005107937,0.0007446257,0.000386588,0.001891357,0.001367705,0.001275648,0.0002423817,0.00006549203],"category_scores_gemma":[0.004726006,0.0004897504,0.0004487408,0.001257733,0.0001528464,0.0005351508,0.0006068427,0.0009565021,0.0000195455],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002522069,"about_ca_system_score_gemma":0.0006553357,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001871491,"about_ca_topic_score_gemma":0.00003362893,"domain_scores_codex":[0.9927923,0.003323362,0.0008511942,0.001582069,0.000431734,0.001019319],"domain_scores_gemma":[0.9950221,0.002942824,0.0003460914,0.001034257,0.0004486647,0.0002060512],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00004955139,0.0001077833,0.0002350653,0.0001322169,0.0002019074,0.000002689561,0.006505023,0.004698807,0.001004657,0.04040192,0.008412892,0.9382475],"study_design_scores_gemma":[0.0007846784,0.0005162414,0.00007990016,0.0002697248,0.0001539942,0.000003940689,0.0002477019,0.6826953,0.004259313,0.001580153,0.3089782,0.0004308048],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0005507849,0.001154187,0.9771551,0.01190447,0.003443184,0.0009887726,0.00000457556,0.0005400318,0.00425884],"genre_scores_gemma":[0.03120816,0.00001822175,0.9221755,0.003773271,0.0002673579,0.0001589501,0.000005256929,0.00003241449,0.04236089],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9378167,"threshold_uncertainty_score":0.9997554,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1356687935760371,"score_gpt":0.4992251913451128,"score_spread":0.3635563977690757,"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."}}