{"id":"W3164193251","doi":"10.48550/arxiv.2105.11698","title":"Guiding the Growth: Difficulty-Controllable Question Generation through Step-by-Step Rewriting","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Rewriting; Interpretability; Controllability; Computer science; Task (project management); Inference; Artificial intelligence; Question answering; Rule of inference; Programming language; Mathematics","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"],"consensus_categories":[],"category_scores_codex":[0.0005680108,0.0003258107,0.0003327778,0.00007813855,0.0005079433,0.0006012877,0.001557771,0.0002981111,0.00001376947],"category_scores_gemma":[0.0001443802,0.0003161692,0.0001980793,0.0004549163,0.00005270961,0.0008634682,0.001630639,0.0004771583,0.00001638786],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003472107,"about_ca_system_score_gemma":0.0001691968,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001118403,"about_ca_topic_score_gemma":0.0001487242,"domain_scores_codex":[0.997331,0.0004402676,0.0003365417,0.001303978,0.000184189,0.000403989],"domain_scores_gemma":[0.9977524,0.0001436662,0.0003400019,0.001292246,0.0003907465,0.00008089668],"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":[0.00001424746,0.000120746,0.001077959,0.0001296163,0.0002038899,0.0002558765,0.000924345,0.4527656,0.006346128,0.5291182,0.006826547,0.002216871],"study_design_scores_gemma":[0.0004039572,0.00001510818,0.0000526993,0.0001396633,0.00006030555,0.000007116302,0.0002649549,0.9947765,0.00106743,0.002068668,0.0007730325,0.000370512],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1416094,0.0003758545,0.8543162,0.0006760583,0.0008668121,0.0002912453,0.00000493747,0.0002298247,0.001629606],"genre_scores_gemma":[0.9742386,0.0003253107,0.02356092,0.0003249857,0.0003771078,0.000003165238,0.00004729,0.0000202186,0.001102387],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8326292,"threshold_uncertainty_score":0.999929,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09930973333867954,"score_gpt":0.2001857070874032,"score_spread":0.1008759737487236,"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."}}