{"id":"W4385571841","doi":"10.18653/v1/2023.acl-long.758","title":"Learning New Skills after Deployment: Improving open-domain internet-driven dialogue with human feedback","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Software deployment; Computer science; The Internet; Domain (mathematical analysis); Artificial intelligence; Machine learning; World Wide Web; Software engineering","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":[0.0002818579,0.0001877192,0.0001988029,0.0001324278,0.000125703,0.0006746955,0.001504951,0.00005629157,0.0001061589],"category_scores_gemma":[0.00001818659,0.0001524346,0.00004453339,0.0003870195,0.00002564525,0.0007613235,0.001923019,0.0002331806,0.0003866207],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007376979,"about_ca_system_score_gemma":0.0000800355,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001334411,"about_ca_topic_score_gemma":0.0007871487,"domain_scores_codex":[0.998311,0.00007122186,0.0002522963,0.0006389372,0.0002671463,0.0004594274],"domain_scores_gemma":[0.9991058,0.00005215221,0.00008298588,0.0005565092,0.00003261662,0.0001699869],"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":[0.0001107446,0.0003928321,0.1574056,0.0001825557,0.0003572553,0.001350244,0.05816798,0.02829771,0.01357819,0.1574374,0.01378008,0.5689394],"study_design_scores_gemma":[0.008350652,0.001976805,0.06429815,0.0008085129,0.00005826433,0.0001198491,0.001191622,0.8638551,0.007778598,0.02084164,0.02747045,0.003250382],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3465692,0.000009290804,0.6480778,0.0005284537,0.0001538069,0.0002666968,2.70579e-7,0.0005415896,0.003852805],"genre_scores_gemma":[0.8899207,0.000001793347,0.08698973,0.0003919018,0.0001155214,0.00003683735,0.000004591736,0.00002869449,0.02251025],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8355573,"threshold_uncertainty_score":0.6506104,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01701245785068708,"score_gpt":0.2501860809956955,"score_spread":0.2331736231450084,"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."}}