{"id":"W6891723697","doi":"10.48448/rqwd-cy29","title":"Few-shot Fine-tuning vs. In-context Learning: A Fair Comparison and Evaluation","year":2023,"lang":"en","type":"other","venue":"Open MIND","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Set (abstract data type); Identification (biology); Process (computing); Quality (philosophy); Term (time)","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","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001548824,0.0002612131,0.0004511245,0.0004060182,0.0001124557,0.0007018411,0.0008800517,0.0002036626,0.00245825],"category_scores_gemma":[0.000260374,0.0002655704,0.00004273708,0.0004766091,0.00004673715,0.0003136937,0.0005655499,0.0004962826,0.001218774],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007932633,"about_ca_system_score_gemma":0.0001576941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003545471,"about_ca_topic_score_gemma":0.003969993,"domain_scores_codex":[0.9978082,0.0003782778,0.0003454536,0.0007189016,0.0004508961,0.0002982387],"domain_scores_gemma":[0.9989287,0.0002058568,0.0003475006,0.0003703368,0.0000454078,0.0001022452],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001789156,0.00002858136,0.002209367,0.00001589905,0.0000321118,0.00002126078,0.005405397,0.0008112596,0.00001998016,0.0002545239,0.01889847,0.9722853],"study_design_scores_gemma":[0.00116602,0.0001064635,0.001450949,0.0003613505,0.00001945649,0.000004818198,0.0007742455,0.2106903,0.00001933327,0.00003240026,0.7850164,0.0003581927],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.002966994,0.0009341587,0.01263447,0.0006151779,0.0007367852,0.001508798,0.000004443722,0.0001087745,0.9804904],"genre_scores_gemma":[0.2377042,0.00007646342,0.01802731,0.0001003303,0.0001653062,0.0001185495,0.00006676554,0.000444435,0.7432966],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.971927,"threshold_uncertainty_score":0.9999797,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1223040697787293,"score_gpt":0.3728091254878406,"score_spread":0.2505050557091113,"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."}}