{"id":"W4414622738","doi":"10.1007/s10664-025-10731-0","title":"DeepCodeProbe: Evaluating Code Representation Quality in Models Trained on Code","year":2025,"lang":"en","type":"article","venue":"Empirical Software Engineering","topic":"Software Engineering Research","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"Huawei Technologies (Canada); Polytechnique Montréal","funders":"Fonds de Recherche du Québec - Santé; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Representation (politics); Software quality; Code (set theory); Quality (philosophy); Code smell; Source code; Software; Static program analysis","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":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001524459,0.0003026492,0.0003992014,0.0005759531,0.00009017318,0.0001633106,0.0009931377,0.0001774838,0.000009456379],"category_scores_gemma":[0.009862482,0.0003237169,0.0001203411,0.002037353,0.00002492132,0.000496674,0.0003398784,0.0006770328,0.00002209077],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004908887,"about_ca_system_score_gemma":0.0001600621,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004447435,"about_ca_topic_score_gemma":0.00001369882,"domain_scores_codex":[0.9968296,0.0001665782,0.000640987,0.0008870212,0.0007875194,0.0006883411],"domain_scores_gemma":[0.9944384,0.004256038,0.00005987357,0.0009540233,0.000129729,0.0001619734],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002270549,0.00009600682,0.05945088,0.0001019532,0.00002529615,0.00002484076,0.0005447614,0.9269376,0.0003252631,0.003317815,0.0003016361,0.008851197],"study_design_scores_gemma":[0.0008118917,0.00007099273,0.1946229,0.0001895208,0.000004159376,0.000003082802,0.00001080292,0.7998773,0.001033949,0.002805253,0.0002237955,0.000346379],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2459962,0.0001226148,0.751425,0.0006124289,0.0003258441,0.0003172807,0.000005283383,0.001106594,0.00008871525],"genre_scores_gemma":[0.880935,0.000007008774,0.1183241,0.0002736243,0.00006034018,0.0001705617,0.000009481593,0.00003651788,0.0001834122],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6349387,"threshold_uncertainty_score":0.9999215,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1192894051877803,"score_gpt":0.4194135589370064,"score_spread":0.3001241537492261,"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."}}