{"id":"W4389523869","doi":"10.18653/v1/2023.conll-babylm.5","title":"Grammar induction pretraining for language modeling in low resource contexts","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University; Mila - Quebec Artificial Intelligence Institute","funders":"","keywords":"Baseline (sea); Computer science; Hyperparameter; Grammar; Context (archaeology); Natural language processing; Language model; Artificial intelligence; Resource (disambiguation); 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":[],"consensus_categories":[],"category_scores_codex":[0.0005488825,0.00007207256,0.00009475079,0.0001715142,0.00005381877,0.00006702248,0.0003150269,0.0000569344,0.000002847869],"category_scores_gemma":[0.00009633869,0.00007056847,0.00003379074,0.0003634036,0.000006099909,0.0002567362,0.0001061927,0.00009961698,0.00001314778],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003232143,"about_ca_system_score_gemma":0.00002417361,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006701157,"about_ca_topic_score_gemma":0.00005642577,"domain_scores_codex":[0.9990984,0.00002497176,0.0001851154,0.0002991499,0.0001315307,0.0002608946],"domain_scores_gemma":[0.9995602,0.00006996193,0.0000257406,0.00028321,0.00002319041,0.00003774502],"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":[0.00001062119,0.00002270277,0.0001462552,0.00006044252,0.000008713955,0.00001792516,0.01945566,0.349638,0.002358615,0.04156645,0.000340786,0.5863739],"study_design_scores_gemma":[0.0002588773,0.00001363207,0.00002245313,0.00003344646,6.662455e-7,0.00000187025,0.0008284597,0.9954987,0.0003815208,0.002757652,0.0001153866,0.00008738413],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2583456,0.00002059907,0.7398043,0.0004344732,0.0001274793,0.000154519,4.197338e-7,0.0003064866,0.0008060901],"genre_scores_gemma":[0.952441,0.000001293507,0.04662992,0.0001536337,0.00008608655,0.00003233883,0.000004064577,0.000008678067,0.0006429329],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6940954,"threshold_uncertainty_score":0.2877699,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05666310754084649,"score_gpt":0.2865182320582126,"score_spread":0.2298551245173661,"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."}}