{"id":"W4389524398","doi":"10.18653/v1/2023.emnlp-main.63","title":"Tree of Clarifications: Answering Ambiguous Questions with Retrieval-Augmented Large Language Models","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"Ministry of Science and ICT, South Korea; National Research Foundation of Korea; Institute for Information and Communications Technology Promotion; Electronics and Telecommunications Research Institute; National Research Foundation","keywords":"Computer science; Question answering; Ambiguity; Tree (set theory); Set (abstract data type); Artificial intelligence; Code (set theory); Natural language processing; Language model; Information retrieval; Domain (mathematical analysis); Machine learning; Mathematics; Programming language","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":[0.000271596,0.00008165307,0.0001126081,0.0001583433,0.00007590469,0.00004147836,0.0004326224,0.0000443291,0.00001277495],"category_scores_gemma":[0.00002766613,0.00006868359,0.00003010649,0.0008406918,0.0000175923,0.0003522416,0.0001215063,0.00009237971,0.00002422478],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001877345,"about_ca_system_score_gemma":0.00005246756,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001951014,"about_ca_topic_score_gemma":0.00009378251,"domain_scores_codex":[0.9990518,0.00003053696,0.000199365,0.000267088,0.000238306,0.0002129055],"domain_scores_gemma":[0.9991127,0.00002990613,0.00005681628,0.0006689095,0.00008038386,0.00005129744],"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":[0.00002486754,0.0002532077,0.0005174311,0.00008842412,0.00009052242,0.00005960508,0.009060771,0.0695298,0.02414752,0.8650732,0.000583509,0.03057114],"study_design_scores_gemma":[0.000271664,0.00003320836,0.001047244,0.00003352579,0.000005658984,0.000004311993,0.0003910981,0.993055,0.002668909,0.002318139,0.0000683032,0.0001029583],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06546577,0.00006293841,0.9286313,0.0005594869,0.00005526549,0.0001276722,0.000004844466,0.0005392497,0.004553457],"genre_scores_gemma":[0.933526,0.00001779813,0.0645825,0.00007626792,0.00001613407,0.00001017111,0.000007402474,0.000008784854,0.001754971],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9235252,"threshold_uncertainty_score":0.2800836,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02744474896684697,"score_gpt":0.2701363584174316,"score_spread":0.2426916094505846,"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."}}