{"id":"W4412889455","doi":"10.18653/v1/2025.acl-srw.5","title":"Rethinking Full Finetuning from Pretraining Checkpoints in Active Learning for African Languages","year":2025,"lang":"en","type":"article","venue":"","topic":"Intelligent Tutoring Systems and Adaptive Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Canadian Institute for Advanced Research; Nvidia","keywords":"Computer science; Natural language processing; Artificial intelligence; Programming language","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002704356,0.001142809,0.0009253747,0.0006072683,0.000622704,0.001416238,0.002553423,0.001193772,0.002314509],"category_scores_gemma":[0.009236842,0.0006151355,0.0005435979,0.0004185115,0.001075261,0.002992199,0.002345855,0.002334305,0.0008146616],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000713893,"about_ca_system_score_gemma":0.001270514,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00741923,"about_ca_topic_score_gemma":0.01314502,"domain_scores_codex":[0.9991899,0.0002856535,0.0000489179,0.0002699681,0.00009267153,0.000112848],"domain_scores_gemma":[0.9956566,0.003089101,0.0001872312,0.0005505015,0.0003651236,0.0001515689],"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":[0.0006381237,0.0005885502,0.01051623,0.0002369758,0.0001550596,0.0001619006,0.001079328,0.3939623,0.02449662,0.002616978,0.002213703,0.5633342],"study_design_scores_gemma":[0.00005260935,0.0001889074,0.001099776,0.00003093682,0.00003142068,0.00005060733,0.0002389652,0.9804214,0.01207095,0.004313251,0.001476058,0.0000251481],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3117842,0.0007593455,0.6708229,0.0006193588,0.0001206518,0.0002470519,0.0002228317,0.01137702,0.004046681],"genre_scores_gemma":[0.8630671,0.0001077348,0.133919,0.0002757116,0.00002444293,0.0002241159,0.0002847139,0.0004624449,0.001634612],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00741923,"threshold_uncertainty_score":0.01475209,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02508123280539928,"score_gpt":0.2826568588096516,"score_spread":0.2575756260042523,"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."}}