{"id":"W3207166518","doi":"10.18653/v1/2022.naacl-main.341","title":"Symbolic Knowledge Distillation: from General Language Models to Commonsense Models","year":2022,"lang":"en","type":"article","venue":"Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","topic":"Topic Modeling","field":"Computer Science","cited_by":150,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Naval Information Warfare Center Pacific; Natural Sciences and Engineering Research Council of Canada; Defense Advanced Research Projects Agency; Allen Institute for Artificial Intelligence","keywords":"Computer science; Computational linguistics; Language model; Artificial intelligence; Natural language processing; Cognitive science; Linguistics; Philosophy; Psychology","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003484638,0.0001866791,0.0003752212,0.0001405285,0.0004403511,0.00004940245,0.002957994,0.00004133933,0.000002932397],"category_scores_gemma":[0.0009644854,0.0001342728,0.0002357264,0.000581282,0.0002177819,0.00007500238,0.002281921,0.0002774492,3.570841e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001891339,"about_ca_system_score_gemma":0.00009024014,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004301054,"about_ca_topic_score_gemma":0.0001129528,"domain_scores_codex":[0.9981584,0.00003884526,0.0005439167,0.0003326847,0.0007040136,0.0002221632],"domain_scores_gemma":[0.9967321,0.0002557832,0.001495996,0.0004848553,0.00100642,0.00002489835],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002291805,0.00008964619,0.01303906,0.0000461957,0.0001437821,1.521142e-7,0.01318386,0.2067942,0.001279672,0.7620014,0.0002374982,0.00316163],"study_design_scores_gemma":[0.0004390065,0.0001342307,0.01021671,0.00007803184,0.0001036115,0.000001732678,0.005089923,0.7253549,0.003852173,0.254204,0.0001803143,0.0003454046],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9808891,0.0001241132,0.01428857,0.001546734,0.0004200007,0.0006863624,0.0005010008,0.0001680277,0.001376068],"genre_scores_gemma":[0.9863427,0.000001937957,0.01315339,0.00007935718,0.00006539603,0.00005644086,0.000009781686,0.00001610274,0.0002748976],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5185606,"threshold_uncertainty_score":0.5496739,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02968002282188094,"score_gpt":0.2597489079193148,"score_spread":0.2300688850974339,"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."}}