{"id":"W3160563524","doi":"10.18653/v1/2021.acl-long.114","title":"Reflective Decoding: Beyond Unidirectional Generation with Off-the-Shelf Language Models","year":2021,"lang":"en","type":"preprint","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Army Research Office; Naval Information Warfare Center Pacific; Natural Sciences and Engineering Research Council of Canada; Defense Advanced Research Projects Agency; Allen Institute for Artificial Intelligence","keywords":"Decoding methods; Computer science; Contextualization; Context (archaeology); Sentence; Task (project management); Artificial intelligence; Reflection (computer programming); Unsupervised learning; Natural language processing; Machine learning; Algorithm","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":[0.0003117598,0.0002318235,0.000213228,0.0001322882,0.0001955635,0.0005315733,0.0007903884,0.0001614399,0.00003049616],"category_scores_gemma":[0.00001906608,0.0001670306,0.00008651279,0.0002808042,0.0000289374,0.0003835305,0.0009749621,0.0004978199,0.000004111751],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001764804,"about_ca_system_score_gemma":0.0004601982,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000228819,"about_ca_topic_score_gemma":0.0005488833,"domain_scores_codex":[0.9981292,0.00012608,0.0002414176,0.0008254604,0.00045184,0.0002259358],"domain_scores_gemma":[0.9984565,0.00006361405,0.0001267619,0.001027662,0.0002615457,0.00006394223],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000007755509,0.00007193784,0.00003432929,0.00003249986,0.0001872008,0.00006321503,0.009297845,0.7046842,0.001657783,0.225128,0.0006063714,0.05822886],"study_design_scores_gemma":[0.0001122714,0.00001830087,0.00001099853,0.00002536646,0.00001471478,0.00005108862,0.0001800413,0.9866508,0.005076078,0.007531628,0.0000933077,0.0002353511],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02806939,0.0006275322,0.9382516,0.001173488,0.0007064316,0.0002400218,0.00000255949,0.0002155011,0.03071354],"genre_scores_gemma":[0.5579488,0.00004482555,0.4383785,0.0006560117,0.0004159142,0.00007922968,0.00004129034,0.00001819377,0.002417285],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5298793,"threshold_uncertainty_score":0.681131,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04713021984422529,"score_gpt":0.2817480599094048,"score_spread":0.2346178400651795,"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."}}