{"id":"W7077860310","doi":"10.48448/6za3-aa95","title":"ALPACA AGAINST VICUNA: Using LLMs to Uncover Memorization of LLMs","year":2025,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Windsor; Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Memorization; Training (meteorology); Training set; Base (topology); Iterative learning control; Factor (programming language)","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"],"consensus_categories":[],"category_scores_codex":[0.0005969316,0.0002759184,0.0003453346,0.0006314713,0.0001696967,0.0001309447,0.002104778,0.0002070321,0.0001477859],"category_scores_gemma":[0.0004923512,0.000254742,0.0000606545,0.002512211,0.0003156676,0.0002312195,0.0009462921,0.0001812981,0.00003692761],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001168784,"about_ca_system_score_gemma":0.001089714,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001682239,"about_ca_topic_score_gemma":0.00003105449,"domain_scores_codex":[0.9975911,0.00003311497,0.0003621595,0.0008593176,0.0006939718,0.0004602953],"domain_scores_gemma":[0.9981691,0.00006234799,0.000316173,0.000949704,0.000353029,0.0001496159],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002381634,0.0007646974,0.001088404,0.001539233,0.0002329736,0.0001093653,0.001711355,0.1258608,0.157489,0.3194402,0.2352705,0.1564697],"study_design_scores_gemma":[0.0006668813,0.0001578782,0.0001250223,0.001770163,0.00005157538,0.00001753504,0.0002081855,0.3609614,0.04774291,0.01257546,0.574182,0.001541001],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.0001381089,0.00007165051,0.4499311,0.0007030062,0.0009745146,0.0002521793,0.00001458996,0.0001639848,0.5477508],"genre_scores_gemma":[0.1074956,0.00007265852,0.3155789,0.002595321,0.0005880504,0.00001642809,0.0000524666,0.00006717472,0.5735334],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.3389115,"threshold_uncertainty_score":0.9999905,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01544067078710087,"score_gpt":0.269653776102785,"score_spread":0.2542131053156841,"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."}}